Antibiotics and microbial resistance in Brazilian soils under manure application

2.1 Study area

Soils sampling was performed in the Guaporé watershed (Rio Grande do Sul, Brazil), which extends over 2,030 km² of rivers, streams, and lakes, representing 0.57% of its surface. The climate of the catchment area is subtropical superhumid mesothermal with an average precipitation of 1,401 to 2,005 mm yr⁻¹ and an average temperature of 18.3 °C; there is no distinct dry season (Bastos, 2017).

The region is mainly rural, with agriculture as the main human activity. Crop fields and grasslands cover about 32.5% (small cultivated areas) and 9.7%, respectively, of the area. The remaining area is occupied by forests (56.7%; Bastos, 2017). Pastures are mainly cultivated for animal feeding (poultry, pig and dairy cattle farming) and grain crop production (wheat, soybean, and maize). Urban areas cover less than 1% of the total surface area of the watershed.

In the upper part of the watershed, no-tillage systems are commonly used on large farms for soy and maize cultivation, with low livestock numbers. The middle and lower parts of the watershed are characterised by a sloping relief, with a large number of small tobacco farms. Farmers also raise poultry or pigs. This region is characterised by severe soil erosion, mainly in the hilly areas, along with high sediment loss (Tiecher, Caner, Minella, & dos Santos, 2015).

Soil vulnerability is rarely considered in cropping systems (without crop rotation or winter cover crops), resulting in more favourable conditions for erosion.

2.2 Soil collection

In the four subwatersheds Lajeado-Carazinho, Marau, Capingui, and Arvorezinha, we sampled 15 sites (Figure 1). At all sites, soils are regularly amended with the application of animal manure (except the forest sites) or are subject to grazing. Manure application methods differed between sites (Tables 1 and 2).

Pig manure is applied in the subwatersheds of Arvorezinha and Lajeado-Carazinho (SWI1 to SWI4). These sites are cultivated under conventional tillage, with two to three grain crops annually, which may be interrupted by animal grazing periods.

Site SWI1 is reserved for tobacco growth and has been amended with pig manure for 8 years. Site SWI2 is also used for tobacco growth in summer and serves as a pasture in winter; it has been receiving manure for 11 years.

Sites SWI3 and SWI4 are used for grain and silage maize production in summer and for forage production in winter. Site SWI3 has been amended with manure every 2 years for 12 years, whereas SWI4 has been amended for 20 years.

Poultry litter is largely used in the Marau subwatershed (PoL1 and PoL2). These two sites are subject to cropping management with conventional tillage and to annual animal grazing. Sites PoL1 and PoL2 have been amended with poultry litter application for 5 and 6 years, respectively.

Site PoL3, located in the Arvorezinha subwatershed, has also been receiving poultry litter for 8 years.

Sites DaC1, DaC2, and DaC3 correspond to sites used for dairy cattle over periods of 5, 15, and 40 years, respectively, and are continuously fertilised.

As reference soils, several forest soils were sampled. Forest Site Ft1 is situated in the southeastern part of the Arvorezinha subwatershed catchment, whereas Forest Site Ft2 is situated in the eastern part of the Marau subwatershed. FLONA corresponds to the national forest (FLONA—Federal Conservation Unit), situated in the Capingui subwatershed. In the 1940s, the FLONA site was demarcated, resulting in 1,300 ha of forest, including 450 ha of native Atlantic Forest.

The restored forest (Rf) sites are sites in which ecological processes have been re-established, resulting in increased biodiversity levels and functioning. These areas have previously been used for agricultural activities but have not been fertilised for the last 50 years.

Sampling was performed at a depth of 0- to 20-cm layer with an auger. The samples were dried and subsequently prepared for chemical analyses.

2.3 Analysis of pharmaceuticals

Twelve pharmaceuticals (11 antibiotics and diclofenac [DFC]) were measured via liquid chromatography–tandem mass spectrometry (Table 3). The antibiotics belonged to the most commonly used ones on family farms in Rio Grande do Sul: macrolide (erythromycin, roxithromycin [ROXY], and tylosin [TYL]), quinolone (norfloxacin [NORFL], ciprofloxacin [CIPRO], enrofloxacin [ENRO], and levofloxacin [LEVO]), sulfonamide (sulfamethazine [SMTZ], sulfaquinoxaline, and sulfamethoxazole [SMX]), and tetracycline (oxytetracycline [OXYT]).

For analysis, 500 mg of soil was extracted by pressurised solvent extraction (ASE™350, Thermo Fisher Scientific, Waltham, USA) at 80 °C, using methanol/water (1:2; vol/vol) as extraction solvent. The extracted volume (approx. 30 ml) was then adjusted to 500 ml with deionised water. The final solution contained <5% of methanol. The extracts were purified by solid-phase extraction (Autotrace™ 150, Thermo Scientific, Waltham, USA) using Oasis^® HLB cartridges (6 cc, 200 mg of sorbent; Waters, Milford, USA), with methanol as eluent. The final extracts were evaporated under mild nitrogen steam until they reached a volume of 100 μl for posterior restitution to 500 μl with methanol/water (10:90; vol/vol). Each soil studied was analysed twice with triplicate measurements.

The samples were finally separated by high-pressure liquid chromatography in an Acquity UPLC^®BEH C18 column (2.1 × 100 mm, 1.7 μm; water) with methanol/water acidified with 0.3% formic acid. Liquid chromatography was coupled to a Q-Exactive Orbitrap™ mass spectrometer (Thermo Fisher Scientific), combining high-performance quadrupole precursor selection with high-resolution/accurate-mass detection. All pharmaceuticals were detected using an electrospray ion source operating in positive mode. Data acquisition and processing were performed using Xcalibur 2.2 software (Thermo Fisher Scientific). Quantification was performed according to the external standard methods. Separation and mass parameters are provided in the Supporting Information.

The identity of each molecule was confirmed by the correspondence of three criteria between the mass spectrum of the analysed compound and the mass spectrum of a certified standard solution of this compound: the exact mass of the compounds (nb. measured in high resolution, R = 20,000 with the Q-exactive mass detector), the retention time (nb. determined with the liquid chromatography system in the given condition of the analytical run), and the isotope pattern of the mass spectra.

Detection limits were calculated on the basis of the results obtained for “matrix samples” calibration (quantity and confidence range) and blanks. The calculation is based on a standard method (Neuilly, 1998). Corrected variance (Sc) is needed to apply this method. The detection limit (LD) can be calculated as follows:

where k is the slope of the calibration curve, and

where ; x_i is the standard concentration; is the standard mean; n is the number of standards; n_b is the number of blanks; Q is the calibration residue; and Q_b is the residue in blanks.

It was possible to obtain the most accurate detection limit with four calibrations. The standard curve method was used, with concentrations ranging from 5 to 20 μg/L and with three replicates of injections per point. Ten blanks were analysed.

Recovery experiments were performed only with deuterated standards of sulfamethazine (¹³C₆) and sulfamethoxazole (D₄). In the literature, values between 80% and 120% are considered as satisfying for the quantification method. Recovery values of 89% for sulfamethazine and 77% for sulfamethoxazole are acceptable.

2.4 Quantification of resistance genes

We extracted soil DNA by using a Fast DNA^® SPIN kit for faeces, which efficiently isolates PCR-ready genomic DNA from the samples. Extraction was performed according to the manufacturer's instructions, using FastPrep^®135 (MP Biomedicals, California, USA). Extracted DNA was quantified with a Nanodrop spectrophotometer (Thermo Scientific, Waltham, USA) and stored at −80 °C until analysis. Three antibiotic resistance genes, frequently observed in the environment (Zhang, Zhang, & Fang, 2009), were quantified from the soil DNA extracts: sul1 (sulphonamide resistance), qnrA (quinolone resistance), and erm (erythromycin resistance). Genes were detected by quantitative polymerase chain reaction (qPCR; Realplex Eppendorf) and realised by INRA Transfert Environment (Narbonne, France). We used the kit Mix Express qPCR Super Mix, with premixed Rox Invitrogen. Complementary data about qPCR analyses were not furnished by INRA Transfert Environment.

The 16S rRNA-encoding gene was quantified by the SYBR Green assay, using the universal primers 338F and 518R (Might Cycler^® FastStart DNA Master plus SYBR Green I, Roche Life Science). Subsequently, the 16S rRNA-encoding gene quantities were divided by 4.1 (Hardwick, Stokes, Findlay, Taylor, & Gillings, 2008, Klappenbach, Saxman, Cole, & Schmidt, 2001, Stalder, 2012).

3.1 Impacts of manure type on soil contamination

The highest diversity of antibiotics was found in the soils receiving pig manure (Table 3). Several macrolides (minimum, TYL: 1.6 μg kg⁻¹ maximum, ROXY: 11.50 μg kg⁻¹) and several sulfonamides (maximum, SMTZ: 20.9 μg kg⁻¹ minimum, SMX: 1.1 μg kg⁻¹) were found in the Sites SWI1, SWI2, and SWI3. ROXY and TYL were up to 100 times higher than the values typically observed in Chinese soils receiving pig manure (Hou et al., 2015). Significant concentrations of quinolones were also found in the Sites SWI2, SWI3, and SWI4 (maximum, LEVO: 22.2 μg kg⁻¹ minimum, ENRO: 1.90 μg kg⁻¹). These values are similar to those found in a previous study (Zhou et al., 2013). The concentration of OXYT, which was only found in Site SWI2 (5.90 μg kg⁻¹), was about 20 times higher than that found in a study in the United States (0.30 μg kg⁻¹ Boxall et al., 2003). Diclofenac was detected in all SWI soils (maximum: 15.90 μg kg⁻¹ minimum: 2.80 μg kg⁻¹). The presence of this compound is consistent with its wide use as an anti-inflammatory drug and a painkiller in pigs (Cevc & Blume, 2001) and cattle (Sawaguchi et al., 2016).

The lowest concentrations of pharmaceutical substances were observed for soils receiving poultry litter. Soils receiving pig and dairy cow manure were more highly contaminated. However, Wei et al. (2016) stated that poultry litters are significant sources of pharmaceutical compounds. Poultry manure is of high interest for farmers because of its low water content and high nutrient concentration, enabling a reduction of the application frequency compared with other manures (Bastos, 2017). In the present work, soils receiving poultry litter presented several quinolones (Table 3). In particular, LEVO was found in Site PoL2 and CIPRO in Site PoL1, at concentrations of 4.20 and 1.23 μg kg⁻¹, respectively. Nevertheless, CIPRO was found in lower amounts in these soils than in French soils treated with livestock manure (30 μg kg⁻¹ Pourcher et al., 2014) or in soils in China (53.40 μg kg⁻¹ Hou et al., 2015). Site PoL3 contained a DFC level of 1.30 μg kg⁻¹ which is lower than the levels found in other studies (Hou et al., 2015; Zhou et al., 2013).

In the DaC sites, we found high levels of OXYT and DFC. Site DaC1 also presented SMTZ (1.11 μg kg⁻¹), ENRO (1.20 μg kg⁻¹), SMX (2.40 μg kg⁻¹), sulfaquinoxaline (3.41 μg kg⁻¹), OXYT (9.80 μg kg⁻¹), and DFC (38.3 μg kg⁻¹). The substance LEVO was detected in Sites DaC2 (3.90 μg kg⁻¹) and DaC3 (1.60 μg kg⁻¹). Li et al. (2013) also found OXYT (5.10 μg kg⁻¹), SMTZ (0.46 μg kg⁻¹), NORFL (0.85 μg kg⁻¹), CIPRO (0.53 μg kg⁻¹), ENRO (1.18 μg kg⁻¹), and TYL (0.25 μg kg⁻¹) in Chinese soils receiving concentrated dairy cow manure.

The forest soils (i.e., regenerated and natural forests) were characterised by the presence of several antibiotics. This finding gives cause for serious concern, because it suggests that soils not fertilised with manure could still contain pharmaceuticals (Table 3). Only sulfonamides were not found in these soils. In the FLONA sites, only DFC was found, at a concentration of 1.40 μg kg⁻¹. The Forest Sites Ft1 and Ft2 were marked by high concentrations of LEVO (8.10 μg kg⁻¹) and NORFL (27.80 μg kg⁻¹), respectively. The substances OXYT (2.10 μg kg⁻¹), DFC (2.50 μg kg⁻¹), and ROXY (3.10 μg kg⁻¹) were also found in the Forest Site RFt1. Several quinolones, namely, ENRO (1.10 μg kg⁻¹), NORFL (7.20 μg kg⁻¹), and LEVO (17.30 μg kg⁻¹) were also found in the Forest Site Rf2.

The presence of pharmaceutical compounds, in particular synthetic antibiotics such as norfloxacin, in several of the studied forest sites (FLONA, Ft 1, and Ft2) was surprising at first. However, this presence was validated by the analytical method used in this work, which is supported by various controls to avoid ‘false-positive’ detections. Thus, the identity of each molecule was confirmed by the correspondence of three criteria between the mass spectrum of the analysed compound and the mass spectrum of a certified standard solution of this compound: the exact mass of the compound, its retention time, and the isotope pattern of its mass spectra.

Therefore, the presence of these compounds in the forest soils, which have never received manure fertilisation, suggests a contamination by the surrounding agricultural soils. Three putative ways of contamination can be considered.

The first hypothesis considers the possible transport of the antibiotics through aerosols or dusts generated during manure spreading. The forest canopy would act as a trap for contaminated drops or particles, which finally reach the soil surface during rainfall events. Indeed, aerial transport of antibiotics associated with particulate matter has been reported from cattle feed yards by McEachran et al. (2015). This transport could occur up to 2.2 km. A recent study also shows that dry or wet deposition of atmospheric particles contributes to the dissemination of pharmaceuticals and other contaminants (Ferrey, Hamilton, Backe, & Anderson, 2018). This way of contamination is possible in the studied forest sites soils due to the large presence of soils amended with animal wastes in Rio Grande do Sul.

Another hypothesis considers the transport of antibiotics by runoff during extreme rainfall/storm events, especially in the days following manure application (annual rainfall is 1,695 and 1,765 mm for Ft1 FLONA and Ft2 soils, respectively). Dolliver and Gupta (2008), in a 3-year field study in southwestern Wisconsin, compared runoff losses of antibiotics from the application of liquid hog and solid beef manures under chisel ploughing and no-tillage systems. Their work shows that antibiotics could be transported through runoff, especially during the nongrowing season, and antibiotic levels were generally higher in the no-tillage compared with the chisel plough treatment. Field observations suggest that this way of contamination seems to be possible for the FLONA soil, because this soil is located below a large agricultural field (slope ~5%) For the Sites Rf1 and Rf2 (slope ~5%), which represent small fragmented forest stands surrounded by agricultural fields, this contamination pathway is also possible (Appendix).

Rainfall events favour also soil erosion, which in turn may favour the transport of pharmaceutical compounds attached to soils particles. Bailey, Spielmeyer, Frings, Hamscher, and Schüttrumpf (2015) highlight the displacement of 15 veterinary antibiotics (introduced to German agricultural fields via fertilisation) by runoff and soil erosion. The authors argue that this transport pathway should be studied more intensely to investigate the contamination of river sediments, because some veterinary antibiotics tend to be strongly sorbed to soil particles (Pereira Leal, Ferracciu Alleoni, Tornisielo, & Regitano, 2013). For the three forest soils studied, erosion was observed in the surrounding agricultural soils or next to the road. Tiecher (2017) highlighted that unpaved roads (such as those near the FLONA site) are a major source of eroded soils. Figure 2 shows the roads surrounding the FLONA sampling point. Consequently, erosion of agricultural soils might also contribute to an input of contaminated soil particles into forest soils, especially at the forest edge.

Soils can also be a natural reservoir of antibiotic-producing microorganisms (Madigan, Martinko, Dunlap, & Clark, 2009; Popowska et al., 2012). Nearly 50% of soil actinomycetes species can synthesise natural antibiotics (Topp, 1981). Consequently, the absence of erythromycin and TYL in the forest soils shows no contribution of the natural soil biomass. Nevertheless, antibiotics of natural origin cannot be ruled out to explain the presence of ROXY and OXYT in these soils.

However, it should be noted that this finding is a case-specific result, because the contamination of manure can greatly vary between different farms. In particular, the contamination of animal excreta depends on the use of pharmaceuticals. However, NORFL, which is an antibacterial compound frequently used in pig production in Brazil, was not found in the SW soils. It should be noted that the metabolism of pharmaceuticals in pigs is not that different from the metabolism in cattle; we can therefore not explain why norfloxacin occurred in soil fertilised with cattle manure but not in soil fertilised with pig manure.

The pollution induced by pig manure application could be related to several factors, such as the maintenance of animals in confinement, the technological conditions for pork production (with the use of more pharmaceuticals than in dairy production), the application of manure with a high content of humidity, allowing its mobility in the soil, and the maturation period of manure, which is often neglected.

3.2 Potential transfer of pharmaceuticals by erosion

Particles carrying pharmaceuticals may be transported to rivers during storm events and in soil runoff. The Guaporé watershed presents an erosion rate of 394.6 Mg·km⁻²·yr⁻¹ (Tiecher et al., 2015). Regardless of the soil potential soil particle export, there is a significant risk of transporting these chemicals into water systems and higher trophic levels. Bailey et al. (2015) found the substances SMX (30–37 μg kg⁻¹), SMTZ (>20 μg kg⁻¹), and tetracyclines (<750 μg kg⁻¹) in river sediments adjacent to agricultural fields in Germany. Considering a concentration of 1 μg kg⁻¹ of compound (e.g., SMX), the erosion rate in the Guaporé catchment may lead to a potential export of 394.6 mg of compound per km² per year (assuming that the compound concentration in soil remains stable throughout 1 year, i.e., not considering biodegradation or additional contamination). Considering a concentration of 38.3 μg kg⁻¹ of compound (e.g., DFC), the erosion rate can lead to a potential export of 15.1 g of compound per year. These findings suggest that even if pollutants are retained by the soil and not leached by runoff, erosion may constitute a significant route to bring them into the aquatic environment.

3.3 Resistance genes in soil

All soils receiving poultry litter residue showed a similar number of resistance genes (sul1: 10⁻⁷ genes per bacteria, qnr A: 10⁻⁴ genes per bacteria, and erm: 10⁻⁷ genes per bacteria). The erm gene was not found in the Site PoL2. Thus, compared with slurry from pig and cattle farms, the application of poultry litter seems to limit the transfer of antibiotic resistance. Wei et al. (2016) suggested that the specific composition and the particularities of poultry manure may influence its pollution potential. Furthermore, PoL soils were characterised by low amounts of antibiotics, impeding the presence of resistance genes in bacteria.

Long-term dairy cow grazing (with the application of manure for more than 15 and for 40 years) leads to high concentrations of sul1 genes (10⁻⁶–10⁻⁵ genes per bacteria) in soils. On the contrary, practices involving grazing shorter periods lead to low concentrations (10⁻⁸ genes per bacteria). Significant differences in sul1 contents between these soils were confirmed by Student's t test (p < .05).

Similarly, differences were also found between the different DaC soils, depending on their location and their land use, although these differences were not confirmed by statistical testing. Pastures from the Marau region had higher qnr A and erm gene amounts (DaC2: 10⁻³ and 10⁻⁶ genes per bacteria, respectively; DaC3: 10⁻⁴ and 10⁻⁷ genes per bacteria, respectively) compared with the pastures from the Lajeado-Carazinho region (DaC1: 10⁻⁵ and 10⁻⁸ genes per bacteria, respectively). Some DaC soils are used for animal grazing but with a large dispersion of the manure. Consequently, the amount of resistance genes is limited, although these repeated inputs seem to maintain these resistances in microbial communities.

Sites SWI1 and SWI4 were characterised by low amount of sul1 gene (~10⁻⁷ genes per bacteria). They also presented similar values of quinolone and macrolide resistance genes when compared with Site SWI3. It should be noted that on Site SWI3, manure is only applied every 2 years. Such nonintensive practices are probably related to the low resistance observed. Heuer and Smalla (2007) reported that the presence of antibiotic-resistant bacteria was especially high for pigs compared with other animals. Furthermore, the resistance is linked to the doses of antibiotics applied.

The sul1 gene was not observed in the forest soils, except in the Ft1 site (10⁻⁸ genes per bacteria; Table 4). This gene is widely found in environments exposed to anthropic activities (Wellington et al., 2013). For example, a typical baseline for sul1 genes below 1.5 × 10⁻² genes per bacteria was defined for 21 lake sediments (Czekalski, Sigdel, Birtel, Matthews, & Bürgmann, 2015). Thus, the absence of sul1 genes may be linked to the absence of sulfonamides in the studied forest soils. The Sites FLONA, Ft1, and Ft2 presented lower values of erm (10⁻⁷ genes per bacteria) and qnrA genes (10⁻⁴ genes per bacteria) than the Sites Rf1 and Rf2 (10⁻⁶ and 10⁻³ genes per bacteria, respectively; Table 5). The two restored forests have been cultivated in the past (~50 years ago); our results suggest that manure application in the surrounding fields could have contaminated these sites via aerosol deposition or runoff. Indeed, the Lajeado-Carazinho region produces (with frequent manure application) different crops on hilly areas. Heuer and Smalla (2007) showed that antibiotic resistance conferred by plasmid exchange was more pronounced in microbial communities of preserved soils exposed to manure.

3.4 Specific manure impact

The concentrations of antibiotics and resistance genes were used to discriminate the different soil samples according to their type of management (without application, DaC, SWI, and PoL). The graphical representation of the discriminant analysis displays only a partial differentiation between the management types. Thus, the results show an overlap of the ellipse of confidence for dairy cow and poultry litter soils, for poultry litter and nonfertilised soils, and for soils without pig manure fertilisation. However, this approach also shows that some types of management are different. Hence, the type of management can induce a ‘specific pollution fingerprint,’ but some of these fingerprints can present similitudes. Nevertheless, discriminant analysis points out that soils fertilised with dairy cow manure seem to present specific contents of in sul1 gene and tetracyclines. Soils with or without fertilisation with poultry litter seem to be impacted by the presence of the resistance genes erm and qnr A. Soils treated with pig manure tend to be more impacted by the presence of antibiotics (Figure 3).

Studies agree that the chronic exposure to antibiotics favours the presence of resistance genes in microbial communities (Balcázar, Subirats, & Borrego, 2015; Martinez, 2012). Thus, some correlations are frequently observed between the abundance in resistance genes and the concentration of antibiotics (Gao, Munir, & Xagoraraki, 2012). According to the results of Pearson's correlation test, no correlation was found between the gene contents (sul1, erm, or qnr A) and the antibiotic contents, suggesting that the abundance of resistance genes may be associated with the presence of other antibiotics (not analysed in the present work) or with a direct input of genes via manure application.

3.5 Contamination of Guaporé soils versus contamination in other countries

The concentrations of antibiotics and resistance genes found in the Guaporé soils tend to be similar to those observed in other agricultural soils worldwide (see Bailey et al., 2015; Heuer et al., 2011; Hou et al., 2015; Li et al., 2013; Malik et al., 2008; Popowska et al., 2012; Thiele-Bruhn, 2003; Wei et al., 2016; Zhou et al., 2013). Indeed, animal production in feedlot systems (poultry and pigs) is based on similar technologies. Scientific advances in nutrition and disease control have been rapidly transformed into technologies adopted by almost all animal breeders in all countries (Shields & Orme-Evans, 2015). The system is based on confining large quantities of animals in specifically constructed buildings (Jong & Guémené, 2011; Rotta et al., 2009; Sather, Jones, Schaefer, Colyn, & Robertson, 1997), on a balanced diet (Benchaar, Pomar, & Chiquette, 2001) and with specific schedules to administer all supplements, including veterinary drugs, in order to obtain maximum feed conversion in the shortest time. The south of Brazil presents a large amount of concentrated production areas administrated by family farmers working in monoproduction systems (Tiecher, 2017). In Brazil, as in Canada or the United States, antibiotics are used as growth promotors, whereas in the European Union, the use of antibiotics for such purpose was banned in 2006 (Maron, Smith, & Nachman, 2013). As a consequence, manures with similar or different qualities are used to fertilise soils, depending on the specific laws and regulations.

The seasonal management and the type of the soil receiving the animal waste may vary between countries or regions within a given country. For example, in Brazil, the limits for manure application are related to the zinc and copper quantities, and the volume of the manure applied (m³/ha) is determined by the pH, the C/N ratio, and concentrations of total organic matter, total carbon, phosphorus, and potassium (IAP). In Germany, fertiliser application restrictions already exist to prevent the overland transport of fertiliser-based contaminants. For example, application should not take place in winter or directly before predicted heavy rainfall (Bailey et al., 2015).

However, the mechanisms of transformation/degradation, the adsorption/desorption processes, among others are identical (Thiele-Bruhn, 2003). The intensity of changes in veterinary drugs is certainly different, but the physicochemical or biochemical principles are comparable. Thus, it can be stated that practices and management strategies adopted in the Guaporé watershed are similar to those used in other countries. The values presented in this work are rather a result of soil management strategies than of the levels of pharmaceuticals in manure.