**2. Results**

#### *2.1. Predicted Concentrations of Antimicrobials in WWTPs*

According to sales data for 2018 (published by the Polish National Health Fund (NFZ)), among detected antimicrobials, the highest sale in Poland was noted for clarithromycin (up to 1509 kg/month), sulfamethoxazole (up to 1358 kg/month), ofloxacin (up to 829 kg/month), ciprofloxacin (up to 821 kg/month), and clindamycin (737 kg/month). We compared the predicted load of antimicrobials in the WWTP1 (PLoad) (Poland) with the load calculated based on the measured concentrations of the drugs in the influent (water phase) and in the primary sludge (Load <sup>W</sup>+S) (Table A1, Online Resource). For most of the tested antimicrobials, the Load W+S to PLoad ratio was low (up to 20%) because the high metabolism in the human body results in a lower level of parent compound, e.g., the biotransformation ratio of clindamycin is 85% [30]. Moreover, the measured load of fluoroquinolones (ciprofloxacin and norfloxacin) was very close to that of the predicted load, and unlike other antimicrobials that are primarily present in the water phase, ciprofloxacin and norfloxacin are distributed evenly between water and the primary sludge. For six antimicrobials (erythromycin, metronidazole, oxytetracycline, tetracycline, sulfathiazole, and vancomycin), the Load W+S to PLoad ratio was very high and exceeded 1000% because of the low and very low predicted load value. Note that almost all antimicrobials in Poland are available by prescription and most of them are reimbursed; however, some, such as vancomycin, are primarily used in hospitals, while tetracyclines (oxytetracycline and tetracycline) and sulfathiazole are primarily used in veterinary medicine and are not reported by NFZ.

#### *2.2. Antimicrobial Concentrations in Influents of WWTPs*

The concentrations of 26 antimicrobials in influents from WWTP1 and WWTP2 and receiving waterbodies are shown in Table 1. The significant differences between the concentrations of antimicrobials in the samples collected in di fferent sampling periods were observed. The di fferences were due to seasonal variations in antimicrobial use. According to National Health Fund (NFZ) database, in summer the consumption of antimicrobials was low (Table A1). In autumn and winter, the consumption increased significantly, probably due to numerous infections occurring each year in that period [31].



CFR (270 ng/L), FLRX (3.1 ng/L), LOM (0.9 ng/L), NAL (6.1 ng/L), PEF (12 ng/L), SDD (3.7 ng/L) were not detected. Their MDLs (method detection limits) are provided in parentheses. 1 Presented MDLs and method quantitation limits (MQLs) were calculated for influents. The values for effluents are about two times lower and for surface water about four times lower.

Nineteen and 18 out of 26 analyzed antimicrobials were detected in the wastewater from WWTP1 and WWTP2, respectively. In most cases, the antimicrobial levels were higher in wastewater of WWTP2 than in wastewater of WWTP1. The mean values of antimicrobials' concentration in analyzed samples ranged from <MDL (method detection limit) to 7400 ng/L. In influent from WWTP1, the average concentrations of two antimicrobials, i.e., ciprofloxacin and sulfamethoxazole, were higher than 1000 ng/<sup>L</sup> and concentrations of 10 antimicrobials was higher than 100 ng/L. While in influent from WWTP2, the concentration of four antimicrobials, i.e., ciprofloxacin, metronidazole, sulfamethoxazole, and vancomycin, exceeded the level of 1000 ng/<sup>L</sup> and concentrations of 12 antimicrobials exceeded the level of 100 ng/L. The highest average concentrations of antimicrobials in influent from WWTP2 were recorded for metronidazole (7400 ng/L), followed by ciprofloxacin (4300 ng/L), vancomycin (u 3200 ng/L), and sulfamethoxazole (3000 ng/L). In other countries, the concentration of metronidazole was lower than 1000 ng/L, that of ciprofloxacin was up to 3800 ng/<sup>L</sup> [32] but frequently below 400 ng/<sup>L</sup> [33], and that of sulfamethoxazole was up to 7900 ng/<sup>L</sup> [34]. There are not much data on vancomycin occurrence in the environment, and it has not been detected in very high concentrations to date as it is primarily used intravenously to treat severe infections in hospitals [35]. The differences in the concentrations of antimicrobials in the influent between two WWTPs can be due to various consumption of antimicrobials in the sampling period and various sources of antimicrobials in these regions. As an example, near WWTP1 is a hospital (449 beds), whereas near WWTP2 is a hospital (458 beds), poultry plant, and galenic laboratory.

#### *2.3. Antimicrobial Concentrations in E*ffl*uent of WWTPs*

The concentrations of 26 antimicrobials in effluents from WWTP1 and WWTP2 are shown in Table 1. In effluents from both WWTPs, the highest average concentrations were observed for azithromycin (up to 650 ng/L), sulfamethoxazole (up to 770 ng/L), ciprofloxacin (up to 312 ng/L), and clindamycin (up to 290 ng/L). In effluent from WWTP2, a higher mean concentration of azithromycin but lower of metronidazole was observed. The percentage contribution of the analyzed antimicrobials in influent and effluent of both WWTPs are presented in Figure 1. To summarize, among antimicrobials tested, ciprofloxacin and sulfamethoxazole were the most dominant antimicrobials in influent and effluent from both WWTPs.

Comparing our data from the literature, the results obtained were similar. The concentration of azithromycin was up to 380 ng/<sup>L</sup> [36] in Switzerland, that of sulfamethoxazole was up to 1300 ng/<sup>L</sup> in Spain [21], and that of clindamycin was up to 5 ng/<sup>L</sup> in Australia [32].

#### *2.4. Antimicrobial Concentrations in Receiving Waters*

In order to assess the impact of antimicrobials on the receiving water bodies, samples were taken upstream and downstream from the WWTPs' discharge (see Section 3.1). The concentrations of 26 antimicrobials in receiving water are shown in Table 1. The river upstream of WWTP1 was more polluted with analyzed antimicrobials than WWTP2 upstream river (Table 1). The highest concentration of sulfamethoxazole (average value: 644 ng/L) was detected upstream of WWTP1. While in the upstream of WWTP2, the highest concentration of clarithromycin (average value: 12.2 ng/L) was detected. The wastewater discharge from WWTP2 resulted in a significant increase of antimicrobials' concentration in the river. This fact can be explained by the high concentrations of antimicrobials in the case of effluent from WWTP2 (Table 1). In the case of WWTP1, taking into account all antimicrobials, no statistical differences between upstream and downstream were observed. Despite this, the average concentration of azithromycin in WWTP1 downstream was more than 15 times higher compared to the concentration in the upstream. The concentrations of antimicrobials in the receiving waters depend on the concentration of antimicrobials in effluent, distance of sampling, or the different flow rate. In our study, WWTP2 effluent effect on antimicrobials' concentrations in the receiving surface water was observed.

**Figure 1.** Percentage contribution (%) of tested antimicrobials in influent and effluent of both WWTPs, *n* = 3.

## *2.5. Antimicrobials' Removal E*ffi*ciency*

The removal efficiency (calculated based on equations described in Section 3.3) of tested antimicrobials was similar for both WWTPs (*p* > 0.05). The average removal efficiency was above 50% for 12 and 10 out of 20 detected antimicrobials for WWTP1 and WWTP2, respectively (Figure 2). Among the antimicrobials from the sulfonamide group, four were detected in effluents and their removal rate ranged from 17% to 80% with an average of 52%. Certain inconsistencies exist in literature about sulfonamide removal, as per the review by Le-Minh et al. [37]. Some researchers have reported an effective removal of sulfonamide [38], although others have mentioned the opposite results [39]. Based on the literature data, this fact might be explained by the differences in operational conditions of each WWTP.

The removal efficiency of trimethoprim was 37% in WWTP1 and 76% in WWTP2 (*p* = 0.2) (the treatment process used in the WWTPs are presented in Section 3.1). Regarding fluoroquinolones, three of them, i.e., norfloxacin, ciprofloxacin, and ofloxacin, were detected in influent of both WWTPs with the average removal being above 80%. The plausible mechanism for the removal of fluoroquinolones from water is sorption to sediment because their concentration was detected as very high in the sludge [40–42] and a slow biodegradation rate was reported [43]. The average removal efficiency for fluoroquinolone antimicrobials estimated by Wang et al. [44] was about 50%. Rodayan et al. [45] reported values of around 60%, while for WWTPs in Switzerland, higher removal efficiencies were observed, reaching up to 87% (for norfloxacin) [46]. In the study of Gao et al. [47], the removal efficiencies of WWTPs for fluoroquinolone antimicrobials ranged from 48% to 72%.

**Figure 2.** Removal efficiency (%) of selected antimicrobials in WWTPs calculated based on equations described in Section 3.3.

In this study, a negative removal rate for azithromycin, clindamycin, lincomycin, and thiabendazole was observed (Figure 2). Moreover, for these antimicrobials, a high variance in removal rate was noted, which agrees with previous studies in which certain antimicrobials were reported to be more abundant in effluents than influents [6,8,10,25,33,48].

The negative removal rate has been detected in the literature. For example, Chunhui et al. [49] reported that the removal efficiencies for macrolide antimicrobials in WWTPs was –4%. Similarly, Gao et al. [47] reported that the removal rate of macrolide antimicrobials was higher, and reached from –34% to 69%. The negative removal rate is explained in many ways. First, certain metabolites can return to the parent pharmaceutical during primary and secondary treatment because of glucuronide conjugate hydrolytic cleavage. Second, pharmaceuticals that sorb to organic matter and particles, as well as accumulate in sediments and biofilms, might be resuspended in wastewater during storm water events. Typically, deconjugation or resorption from particles is observed for fluoroquinolones (ciprofloxacin, ofloxacin, and norfloxacin), lincomycin, tetracycline, trimethoprim, and sulfamethoxazole. This issue is particularly relevant to pharmaceuticals that are mainly excreted with bile and feces-like macrolides. During wastewater treatment, they are redissolved, and, therefore, their concentrations in water increase [10,33].

HRT (hydraulic retention time) is one of the major factors influencing the antimicrobial removal efficiency, particularly for compounds that are readily biodegradable and have a low Kd (low tendency to absorb to sludge) [50]. In this study, HRT was 12 h in WWTP1 and 24 h in WWTP2. Other factors influencing the removal efficiency are adsorption coefficients and persistence of the antimicrobials. Lower levels in effluents could be also interpreted as the removal of antimicrobials because of biodegradation and/or chemical and physical transformations, e.g., hydrolysis or sorption to solid matter. Biodegradation/biotransformation and sorption are the two primary mechanisms occurring in the WWTPs. The sorption to sewage sludge is the primary removal mechanism for certain antimicrobials in the samples, e.g., the percentage of the daily load of antimicrobials was >40% for azithromycin, ciprofloxacin, ofloxacin, and oxytetracycline in sludge (Table 2), while the lowest value was observed for sulfamethoxazole and lincomycin. The differences between the sorption of antimicrobials depend on their physicochemical properties, such as charge and lipophilicity, the properties of the sludge-like chemical nature, and other factors such as pH and redox potential [51]. Sorption may occur by hydrophobic interactions with lipophilic cell membranes of the microorganisms or the lipid fractions

of the suspended solids and electrostatic interactions of positively charged groups of chemicals with negatively charged surfaces of microorganisms or other components of the sludge [33]. Figure 3 shows differences between the concentrations of antimicrobials in wastewater and sludge (separation observed by principal component 1 (PC1) accounted for 45% of total variance). Two groups were designated by the various amounts of oxytetracycline and norfloxacin in the samples. The relative percentage estimated for the antimicrobial was significantly higher in the sludge than in the wastewater. In wastewater, the relative percentage of metronidazole, lincomycin, and sulfamethoxazole was higher compared to that in the sludge. Similarly, a higher contribution of metronidazole and erythromycin in wastewater and ciprofloxacin in sludge was observed by Ostman et al. [24].

**Figure 3.** Differences in antimicrobials' concentrations between the wastewater, sludge samples, and leachate collected from WWTP1 and WWTP2. Principal component analysis (PCA) of the all samples (**A**). Variables of a data matrix are represented as arrows. Visualization of the observed relations on the heat map (**B**). ES, excessive sludge; FS, fermented sludge; In, influent; LS, leachate; Out, effluent; PS, primary sludge; RS, residual sludge.

Our results showed that leachate from WWTP1 contained a significant amount of antimicrobials. The mean concentrations of nine antimicrobials were above 100 ng/<sup>L</sup> (Table 1). The concentrations of azithromycin, ciprofloxacin, norfloxacin, ofloxacin, sulfamethoxazole, and vancomycin were comparable to those in the WWTP1 influent. The results obtained sugges<sup>t</sup> that the antimicrobials are strongly sorbed to solids (Table A1). The antimicrobials bound to solids can be leached to the environment, which should be taken into consideration when conducting the risk assessment. In the Table A1, the Kd values for tested antimicrobials are presented. Generally, Kd is used to estimate the mobility and distribution of the pharmaceuticals in the environment [41,52]. Compounds with low Kd are potentially mobile from soil to the water (through leaching or runoff). According to our study, the most mobile antimicrobials are trimethoprim (log Kd = 5.8), sulfamethoxazole (log Kd = 5.1), lincomycin (log Kd = 5.2), clarithromycin (log Kd = 5.7), clindamycin (log Kd = 5.7), and thiabendazole (log Kd = 5.9).

#### *2.6. Antimicrobials' Concentrations in Sewage Sludge from WWTPs*

Fifteen antimicrobials in the primary (WWTP1-PS) and excessive sludge (WWTP1-ES) from WWTP1 as well as in the fermented (WWTP2-FS) and residue sludge (WWTP2-RS) from WWTP2 were detected (Table 2). Results obtained indicated that a part of the antimicrobials were adsorbed to the sludge along the whole technological process.


**Table 2.** Mean and standard deviation (*n* = 3) of target antimicrobial concentration (ng/g dry weight) in sewage sludge samples in two wastewater treatment plants (WWTPs) located in Poland at Silesian (WWTP1) and Warmian-Masurian Voivodship (WWTP2).

CFR (350 ng/g), FLRX (12 ng/g), LOM (7.9 ng/g), NAL (3.3 ng/g), MTZ (25 ng/g), RIF (150 ng/g), ST (30 ng/g), SDD (39 ng/g) were not detected. Their MDLs (method detection limits) are provided in parentheses; MQL, method quantitation limit; WWTP-PS, primary sludge; WWTP-ES, excessive sludge; WWTP-FS, fermented sludge, WWTP-RS residual sludge; 1calculated based to the lowest Kd observed and concentration of antimicrobials in leachate.

In our study the highest concentrations for fluoroquinolones, such as ciprofloxacin (up to 57 μg/g), norfloxacin (up to 9.5 μg/g), ofloxacin (up to 2.0 μg/g), and azithromycin (up to 2.3 μg/g), were determined in the sludge samples (data not presented). In China, the concentration of norfloxacin, ofloxacin, and ciprofloxacin in sludge was determined from all the major provincial cities, ranging ranged from 0.1 to 15.7 μg/g, from 0.3 to 7.9 μg/g, and from 0.1 to 4.7 μg/g, respectively [53]. In Switzerland (up to 3.3 μg/g) and Sweden (up to 4.2 μg/g), the maximum concentration of norfloxacin was lower than that in China and presented in Poland. Furthermore, the maximum level of other fluoroquinolones was higher in Poland than in Switzerland (up to 0.9 μg/g for ciprofloxacin) and Sweden (up to 4.8 μg/g for ciprofloxacin and up to 2.0 μg/g for ofloxacin) [54,55]. The concentration of azithromycin in the activated sludge was up to 0.16 μg/g and 0.056 μg/g in Germany and Switzerland, respectively, and the values were lower compared to those for Poland (up to 0.71 μg/g) [56]. Recently, a study from Germany reported the concentration of azithromycin and ciprofloxacin in sediments to be up to 0.33 μg/g and 0.71 μg/g [41], respectively.

#### *2.7. Antimicrobial Resistance Risk Assessment*

The risk factor of antimicrobial resistance selection (named as antimicrobial resistance risk, defined as preferential outgrowth of antimicrobial-resistant bacteria) in wastewater was high for the following antimicrobials: azithromycin, ciprofloxacin, clarithromycin, metronidazole, and trimethoprim (Figure 4A). As expected, this factor was the highest in influent from both WWTPs. Similar results were obtained by Ostman et al. [24], who observed a high risk of resistance selection in influents in Swedish WWTPs because of the prevalence of ciprofloxacin and metronidazole. No risk for azithromycin was detected in their study.

**Figure 4.** Estimation of risk quotients of resistance selection for the antimicrobials occurring in wastewater (**A**), sludge (**B**), and sludge-amended soil (**C**). Risk quotient below 0.1 indicates minimal risk (green area), between 0.1 and 1 is medium risk (orange area), and over 1 is high risk (red area). ES, excessive sludge; FS, fermented sludge; In, influent; LS, leachate; Out, effluent; PS, primary sludge; RS, residual sludge.

It is interesting, the notable antimicrobial resistance risk was also predicted upstream of both WWTPs. High antimicrobial resistance risk for ciprofloxacin and medium risk for azithromycin, clarithromycin, and clindamycin was observed in the case of WWTP1, while medium antimicrobial resistance risk for ciprofloxacin was noted in the case of WWTP2. Downstream WWTP1, on top of the previous risk, the risk for metronidazole turned medium, whereas the risk for azithromycin rose to high. In the case of WWTP2, the risk for ciprofloxacin rose from medium to high and the medium risk for ofloxacin, clarithromycin, and clindamycin appeared (data not presented).

The antimicrobial resistance risk was high for azithromycin, ciprofloxacin, clarithromycin, norfloxacin, trimethoprim, ofloxacin, and tetracycline in sludge (Figure 4B). Interestingly, residual sludge could pose a risk to the environment, particularly because of the presence of fluoroquinolones and macrolides. Mean PECsoil (predicted environmental concentration in soil) was mainly lower than 1 ng/g and higher concentrations were observed for ciprofloxacin (up to 41 ng/g), norfloxacin (up to 7.8 ng/g), ofloxacin (up to 1.2 ng/g), and azithromycin (up to 2.3 ng/g) (Table 3). When RQ was analyzed, a high risk for soil was noted for ciprofloxacin (Figure 4C). The data on risk assessment on the antimicrobial resistance selection in sludge and sludge-amended soil presented in this paper are the first.

#### *2.8. Environmental Risk Assessment*

The high risk was observed for both cyanobacteria (Figure 5A) and eukaryotic species (Figure 5B) due to azithromycin, clarithromycin and sulfamethoxazole in effluents. Furthermore, medium risk was predicted for chronic exposure of cyanobacteria to ciprofloxacin, erythromycin, norfloxacin, and ofloxacin (in effluents). The same level of risk was evaluated for chronic exposure of eukaryotic organisms to ciprofloxacin, erythromycin, and roxithromycin (in effluents). Verlicchi et al. [33] published a review, which reported that six antimicrobials posed a high environmental risk for eukaryotes, i.e., erythromycin, ofloxacin, sulfamethoxazole, clarithromycin, tetracycline, and azithromycin. Harrabi et al. [6] observed a high risk for ofloxacin, ciprofloxacin, azithromycin, sulfamethoxazole, and trimethoprim for eukaryotic species; however, clarithromycin was not evaluated. In their studies, the predicted no effect concentration (PNEC) was, however, calculated differently compared to our study. The authors obtained PNEC values 1000 times lower than the toxicity values found for the most sensitive eukaryotic species that were assayed [6,33]. In our study, other factors, i.e.1000, 100, 50 or10, were used to calculate PNEC. Thus, RQ values obtained for the same concentrations of antimicrobials were lower.

In the case of upstream of WWTP1, there was high risk for cyanobacteria posed by azithromycin, norfloxacin, and sulfamethoxazole, whereas medium risk was observed for ciprofloxacin and clarithromycin. The risk for eukaryote was also observed but the level of the risk was different for the antimicrobials, e.g., for clarithromycin high risk was estimated, and for azithromycin, ciprofloxacin, and sulfamethoxazole medium risk was noted.

None of the antimicrobials posed high risk from upstream of WWTP2. However, medium risk was posed by azithromycin, clarithromycin for cyanobacteria, and medium risk for eukaryote posed by clarithromycin were observed. The medium risk for cyanobacteria posed by ciprofloxacin, ofloxacin, clarithromycin, and sulfamethoxazole in WWTP2 downstream was noted. The risk of toxicity due to azithromycin content increased from medium to high comparing the upstream and downstream. Two more antimicrobials, i.e., azithromycin and ciprofloxacin, were predicted to cause the medium risk for eukaryote.

The RQ calculated for cyanobacteria exposed to sludge was medium for azithromycin (primary sludge up to 0.36, excessive sludge up to 0.12, fermented sludge up to 0.39) and clarithromycin (primary sludge up to 0.13, fermented sludge up to 0.16). No data on risk assessment on antimicrobial presence in sludge was found in literature to compare with our data.

Antimicrobials are designed to act on prokaryotic organisms; thus, environmental bacteria are more likely to be adversely affected compared to other environmental species such as aquatic invertebrates and vertebrates. However, compared to cyanobacteria, certain microalgae and macrophytes are more sensitive to certain antifolate and quinolone antimicrobials [27].

**Figure 5.** Estimation of risk quotients of chronic exposure of cyanobacteria (**A**) and eukaryotic organisms (**B**) to the antimicrobials occurring in effluents and receiving water. Calculations were based on the measured concentration of antimicrobials in wastewater and PNEC calculated based on [27]. Risk quotient below 0.1 indicates minimal risk (green area), between 0.1 and 1 is medium risk (orange area), and over 1 is high risk (red area).

#### **3. Materials and Methods**

#### *3.1. Description of WWTPs and Sample Collection*

The raw influent, final effluent, leachate, and sludge samples (i.e., primary sludge and excessive sludge (WWTP1) or fermented sludge and residual sludge (WWTP2)) were collected from two Polish WWTPs during the three sampling campaigns: (1) In June/July 2018, (2) in October 2018, and (3) in December 2018. Simultaneously, the receiving river water samples from upstream and downstream (WWTP1, 0.25 km; WWTP2, 1.8 km) of the outlet were collected. Each sample was collected in triplicate. The rivers' average annual flow is 3.5 m<sup>3</sup>/s (WWTP1) and 3.7 m<sup>3</sup>/s (WWTP2).

WWTP1 is located in one of the cities from Metropolitan Association of Upper Silesia, one of the largest urban areas in the EU and the center of Poland's industries, particularly coal and metal,

with a density of 1600 people per km<sup>2</sup> [57], geographical coordinates: N 50◦ 5 35.881, E 19◦ 3 32.202. WWTP2 is located in the forested and agriculture area of Warmian-Masurian Voivodship in the city with a density of 1960 people per km<sup>2</sup> [57], geographical coordinates: N 53◦ 48 46.700, E 20◦ 26 55.800.

In 2018, WWTP1 had an equivalent population of 189,332 inhabitants, the average flow rate was 26,830 m<sup>3</sup>/d, and the plant was operated with a hydraulic retention time (HRT) of ~12 h and a solid retention time (SRT) of 25 days, while WWTP2 served a population equivalent of 250,000 inhabitants, and had a flow rate of 32,130 m<sup>3</sup>/d, HRT of 24 h, and SRT of 87 days. Both WWTPs receive domestic, hospital, and industrial wastewaters and treat wastewater with a mechanical-biological system with elevated removal of nutrients. The biological section of WWTP1 included sequential reactors activated sludge chambers (the system of chambers of di fferent oxygen conditions: anaerobic, anoxic, and aerobic), secondary settling tank, and anoxic chamber. In the case of WWTP2, the biological part included a pre-denitrification chamber, phosphorus removal tank, nitrification/denitrification chambers, and secondary settling tanks. Sewage sludge produced in the WWTP1 was subjected to a thickening process, methane fermentation, and dehydration. The leachate produced during sludge fermentation (WWTP1-FS) was returned to the pumping station and again treated. The sewage sludge from WWTP2 was used as a soil improver. The detailed description and technical parameters of both WWTPs are presented by Buta et al. [58].

All samples were collected in triplicate and placed in the sterile glass bottles in volumes of 1–2 liters. Sludge samples were collected after mechanical concentration (primary sludge, WWTP1-PS), after gravity concentration (excessive sludge, WWTP1-ES), after an open fermentation pool (fermented sludge, WWTP2-FS), and after all processes (residue sludge, WWTP2-RS), i.e., sludge for management. Then the samples were transported to the laboratory on the same day and stored at a temperature of 4 ◦C until analysis.

#### *3.2. Analysis of Antimicrobials' Concentrations*
