Hospital wastewater effluent: hot spot for antibiotic resistant bacteria

Hospital wastewaters are highly complex effluents containing antibiotic compounds, metabolized drugs, disinfectants (Emmanuel et al. 2005), patients' excrement and microorganisms, potentially containing multidrug-resistant (MDR) genes (Chang et al. 2010; Galvin et al. 2010; Chagas et al. 2011). As such, hospital wastewaters are considered as hot spots for antibiotic resistance, generating an environment for the exchange of antibiotic resistance genes. Gram-negative bacteria bearing multiple bla genes (e.g. blaNDM, blaKPC, blaCTX-M, and blaSHV) are increasingly found in hospital wastewaters (Chagas et al. 2011; Zhang et al. 2012). This is consistent with antimicrobial resistance surveillance in Europe 2015 (ECDC 2017), observing an increase in combined resistance to multiple antimicrobial groups, as well as a high proportion of extended spectrum beta-lactamases (ESBL) producing isolates, leaving few treatment alternatives for patients suffering from infections caused by these pathogens. Besides its impact on treatment, frequent resistance in Gram-negative bacteria may lead to an increased use of carbapenems, thus further favouring the emergence and spread of carbapenem-resistant bacteria.

Hospital effluents are discharged as common community wastewater to the wastewater treatment plant (WWTP), with no pre-treatment or cleaning process. After treatment, water is discharged into surfaces waters. Certain pathogenic microorganisms can remain in an aquatic environment for longer periods, creating dissemination routes and environmental reservoirs of antibiotic resistance genes (ARG) (Meirelles-Pereira et al. 2002; Perron et al. 2008). Over the last two decades, the occurrence of antibiotics in water bodies and subsequent development of microbial resistance has come into scientific and public focus as an issue of potential concern (Kümmerer 2009a; Rodríguez-Gil et al. 2010). Hospital wastewaters are one of the sources contributing significantly to the environmental burden of antibiotics and consequently antibiotic resistance (Kümmerer 2004; Zhang et al. 2009; Davies & Davies 2010; Harris et al. 2012). ARGs can be found in almost all environments and they are currently considered as an emerging pollutant and an ecological problem (Kümmerer 2004; Engemann et al. 2008). Environmental ARGs could serve as a reservoir and can be horizontally transferred to human-associated bacteria and thus contribute to antibiotic resistance proliferation (Khan et al. 2013).

The aim of this study was to evaluate the presence of ESBL and carbapenemase producing Enterobacteriaceae in hospital wastewater from one Slovenian and two Austrian hospitals. At the time of sampling, the information about the occurrence of clinical isolates of ESBL and carbapenemase producing Enterobacteriaceae in all three hospitals was obtained.

Samples were taken at main outflows from a Slovenian 259-bed general hospital (General Hospital dr. Jožeta Potrča, Ptuj), an Austrian 44-bed private surgery clinic (Privatklinik Leech) and an Austrian 230-bed private rehabilitation clinic (Privatklinik Lassnitzhöhe). Sampling was performed in two consecutive years, in July 2017 and July 2018. For each sampling site, samples were taken in 0.5 L sterile glass flasks and transferred on ice to the laboratory for microbiological analysis.

Information about resistant clinical isolates from the Ptuj general hospital during the study sampling period was obtained from MBL data collection software, SRC Infonet (version 2019.03). Data were then analysed with software K21, Infonet OLAP (version 3.10.3).

For isolation of ESBL and/or carbapenemases producing Enterobacteriaceae (CPE), 0.5 mL of the original sample was plated on selective agar in two repetitions, 0.1 mL of 2-fold serial dilutions were also plated on selective agars, 4 mL of diluted samples were filtered (cellulose nitrate filter, 0.45 μm pore size, 47 mm diameter) and membranes were placed onto selective agars. ChomID ESBL agar (bioMerieux, France) (screening of extended spectrum beta lactamase producing Enterobacteriaceae) and chromID CARBA SMART agar (bioMerieux, France) (CARB: KPC and metallo-carbapenemase-type CPE; OXA: OXA-48 type CPE) were used for screening of ESBL and carbapenemase producing Enterobacteriaceae. Agar plates were incubated for 24 h at 37 °C. Colonies were counted for the colony forming units (cfu) calculation and assessed as described in the manufacturer's instructions:

• Ec: Escherichia coli

• KESC: Klebsiella, Enterobacter, Serratia, Citrobacter.

To obtain pure cultures, selected colonies growing on chromID agar were transferred to TSA and incubated for 24 h at 37 °C. DNA was extracted by a simple method involving a use of heat (99 °C for 20 min) for extraction. Up to three morphologically identical colonies were transferred into sterile tubes (Eppendorf) with 50 μL double-deionized water and boiled at 99 °C for 20 min. The tubes were immediately transferred to –20 °C for heat shock (Abdulsamad et al. 2009).

Bacteria were identified at the species level by using matrix-assisted laser desorption ionization time-of-flight mass spectrometry (MALDI-TOF) with MALDI Biotyper Smart (Bruker Daltonics, Massachusetts) and samples were analysed immediately. Data acquisition and processing were performed with the MBT Compass (version 4.1) and FlexControl (version 3.4) softwares. The data acquisition mass range was m/z 4,000–17,000 Da.

Isolates were screened for five different β-lactamase genes (blaNDM, blaKPC, blaCTX-M, blaVIM, blaOXA-48) using specific real time polymerase chain reaction (PCR) assays.

For detection of OXA-48, we used commercially available PCR assay from TIB MOLBIOL (LightMix^® Modular OXA-48 Carbapenemase). The assay is based on the amplification of 80 bp long fragments from the blaOXA-48 gene with specific primers and detected with an R6G labelled hydrolysis probe. It was performed according to the manufacturer's instructions on the LightCycler 96 (Roche, Switzerland).

In-house validated real time PCR protocols were employed for the remaining ARGs and were based on the previously published protocols (Birkett et al. 2007; Khosravi & Mihani 2008; Centers for Disease Control and Prevention (CDC) 2011). Primers and probes were synthesized at TIB MOLBIOL (Berlin, Germany) and are presented in Table S1 (Supplementary Material).

Real time PCR for blaCTX-M genes is a multiplex PCR assay that enables detection and genotyping of CTXM producing members of the Enterobacteriaceace with four different hydrolysis probes. Assays were run on Rotor Gene Q (Qiagen, Germany).

The blaVIM gene was detected with real time PCR based on SYBR Green intercalating dye and performed on LightCycler 2.0 (Roche, Switzerland). Melting temperature was included in the analysis to confirm the identity of the amplicon.

PCR detection of blaKPC and blaNDM-1 in a single reaction was carried out with two different hydrolysis probes on LightCycler 2.0 (Roche, Switzerland) according to the described protocol available at CDC.

The average yearly antibiotic consumption data were obtained for all three clinics and, more specifically, the average consumption of beta lactam and carbapenem antibiotics in 2017 and 2018.

Antibiotic resistance is a major public-health problem worldwide where hospitals (and hospital wastewaters) act as a hotspot for antimicrobial-resistant bacteria. The spread of MDR bacteria through hospital wastewater is a valid cause for concern (Tuméo et al. 2008) since it is plausible that MDR bacteria are selected mainly in hospitals and passed into wastewater. Increased antibiotic use and disposal is leading to higher levels of bacterial resistance (Kümmerer 2004).

In the present study, high levels of ESBL, CARB and OXA type producing Enterobacteriaceae were demonstrated in hospital wastewaters from Slovenian and Austrian hospitals in ranges up to 10⁷ cfu/mL which is similar to previously published results (Morris et al. 2016; Le et al. 2016). The highest concentration of ESBL producing Enterobacteriaceae according to ChomID ESBL agar identification was present in wastewater from the Slovenian general hospital, followed by the Austrian private rehabilitation clinic and the Austrian private surgery clinic. The concentration of ESBL producing bacteria did not vary significantly between 2017 and 2018 in the Slovenian general hospital. In the Austrian private rehabilitation clinic, the concentration of ESBL producing bacteria was approximately the same for the KESC group, but increased by one logarithmic scale for E. coli isolates. In the Austrian private surgery clinic the concentration of ESBL isolates increased by approximately one logarithmic scale (Table S2, Supplementary Material).

The highest concentration of carbapenemase producing Enterobacteriaceae (mainly Klebsiella producing carbapenemase and metallo-carbapenemase-type CPE) according to ChomID CARB agar identification was again present in wastewater from the Slovenian general hospital, followed by the Austrian private rehabilitation clinic and the Austrian private surgery clinic. The concentration of CARB producing bacteria did not significantly vary between 2017 and 2018 in the Slovenian general hospital. In the Austrian private rehabilitation clinic, the concentration of CARB producing bacteria decreased by two logarithmic scales for both E. coli and group of KESC, between 2017 and 2018. In the Austrian private surgery clinic the concentration of CARB E. coli increased by two logarithmic scales, but the concentration of CARB KESC bacteria dropped by three logarithmic scales and no CARB from group KESC bacteria were isolated in 2018 (Table S2).

OXA-48-type producing CPE bacteria were present at the lowest concentrations for all sampling sites. The concentrations of screened bacteria were highest in the Slovenian general hospital, but decreased for both E. coli and KESC group by one or two logarithmic scales in one year. In wastewater from the Austrian private rehabilitation clinic, the concentration of OXA-48-type CPE remained approximately the same during both years. In wastewater from the Austrian private surgery clinic, the OXA-48-type E. coli was present in concentrations of 1 × 10⁵ cfu/mL, but it was not detected in year 2018. The OXA-48-type from group of KESC bacteria were not isolated in any of the samples in the Austrian private surgery clinic (Table S2).

The chromogenic medium used in the study enables rapid identification of carbapenem resistant pathogens since it contains antibiotic for the inhibition of other microorganisms and biochemical markers to differentiate species or groups of species using either chromogenic substrates or fermentable carbohydrates with a pH indicator. The phenotypic test identifies the carbapenemase producers in general without any specification over the class of carbapenemase. In contrast to the manufacturer's claims, false positives can occur, therefore molecular characterization is strongly recommended (Bakthavatchalam et al. 2016).

To confirm the resistant pattern, 89 DNA samples were extracted from colonies grown on selective chromogenic media for the screening for ESBL and carbapenemase producing Enterobacteriaceae. Amplicons of different β-lactamase gene families were obtained from 36 out of 89 DNA samples (Table 1). Molecular characterization is the only available tool for discriminating different carbapenemase encoding genes (Bakthavatchalam et al. 2016). Molecular methods are able to determine the exact mechanism conferring carbapenem resistance, which can be especially helpful during outbreak investigations. The primary limitation of nucleic acid amplification techniques is that only known genes can be targeted and thus genes encoding novel carbapenemases will be missed (Lutgring & Limbago 2016). The resistance patterns were not completely in agreement between cultivation methods and amplification of specific genes. This could be due to the fact that selected chromogenic media are not 100% specific and growth of non-resistant bacteria can occur according to the manufacturer's instructions. Furthermore, bacteria can acquire resistance in several ways (expression of efflux pumps, decreased outer membrane permeability or antibiotic target modification) and last but not least bacteria with unknown resistance gene/alleles may be present. It would be interesting to further identify those with shot-gun cloning in expression plasmids and transfect them into sensitive bacterial strains to render them resistant.

Three isolates were identified as multiresistant, bearing the VIM/CTXM and CTXM/OXA resistant genes (samples 1, 13, 25 Ptuj). The dominant ESBL family was CTXM, represented by genotypes CTXM-1 in 14 different isolates (E. coli, Citrobacter freundii, Klebsiella pneumoniae, Pseudomonas mosselii, Bacillus subtilis) and CTXM-9 in five different isolates (C. freundii, Enterococcus faecium, E. coli). This was followed by 10 OXA-48 positive isolates (P. mosselii, Citrobacter amalonaticus, C. freundii, Klebsiella oxytoca) and 10 VIM positive isolates (C. freundii, Enterobacter kobei, K. oxytoca). KPC or NDM carbapenemases were not identified with PCR screening of the isolates (Table 1).

The majority of resistant bacteria were identified from the Ptuj general hospital wastewater where 18 samples were PCR positive. ARG were detected in species C. freundii, E. kobei, E. faecium, E. coli, K. pneumoniae, K. oxytoca and Pseudomonas mosselli. In Laßnitzhöhe rehabilitation clinic wastewater, 11 samples were PCR positive with detection of ARG in species C. amalonaticus, C. freundii, E. coli and K. oxytoca. In the Leech surgery clinic wastewaters, only six samples were PCR positive for E. coli and there was one PCR positive sample for B. subtilis (Table 1).

At the time of sampling, 42 resistant clinical isolates from the Slovenian general hospital were reported in 2017 and 2018 (Table S3). Two cases of ESBL bacterial infections were also recorded in Laßnitzhöhe rehabilitation clinic in July 2017, but none in 2018. In the Leech clinic no resistant isolates were recorded for both years. The number for resistant clinical isolates (Table S3) present the current situation in the hospitals, there is no information that those isolates are directly related to the ones found in the waste waters.

Significant differences in antibiotic resistant profiles between the three clinics can be correlated to the apparent difference in antibiotic consumption (Tables S1–S3). The average beta lactam antibiotic consumption in the years 2017 and 2018 was 52.85 kg in the Slovenian general hospital, while only 3.18 kg of beta lactam antibiotics was used in the Austrian private surgery clinic, and even less in the Austrian rehabilitation clinic. The average carbapenem antibiotic consumption in the years 2017 and 2018 was 1.74 kg in the Slovenian general hospital, while those classes of antibiotics were not prescribed in the Austrian private clinics. This finding strongly supports the finding of other authors, that the geographic trends in antibiotic resistance are comparable to national or local antimicrobial use levels (Frost et al. 2019).

Where outbreak cases could not be linked to an epidemiological source, the question of unidentified bacterial reservoirs either within the hospital or in the community is increasing. Few outbreaks with environmental transmission of CPE have been previously described in Australia, Spain and Norway (Kotsanas et al. 2013; Tofteland et al. 2013; Vergara-López et al. 2013); the epidemic clone was identified from the siphon of the sink in a room occupied by a K. pneumoniae OXA-48-colonised patient. In Australia there was a report of an outbreak of CRE in an ICU with identical organisms isolated from patients and an environmental source (sinks). Molecular typing confirmed that clinical and environmental isolates were related. Such studies remind us to consider environmental reservoirs as a source of CPE transmission (Clarivet et al. 2016) and highlights the importance of the identification of potential environmental reservoirs for the control of multiresistant bacteria outbreaks (Kotsanas et al. 2013).

Antibiotic resistance pollution is of concern because of the potential horizontal gene transfer into clinically important pathogens (Martínez 2008; Kraemer et al. 2019). Although some ARB can be removed through conventional wastewater treatment processes (Guardabassi et al. 2002; Martins da Costa et al. 2006), conventional wastewater treatment is not sufficient for the removal of hygienically relevant bacteria and achieves only limited reductions (Hembach et al. 2019). The concentrations of antibiotic resistant bacteria in WWTP discharge are 2–4 log scales lower than concentrations in hospital effluents (Hocquet et al. 2016). There are still large numbers (up to 10⁴ cfu/mL) (Hocquet et al. 2016) that survive in the effluent (Pruden et al. 2006; Hembach et al. 2019), therefore resistant bacteria can be found in environmental aquatic systems. Moreover, the comparison of resistant profiles from clinical and environmental isolates from WWTPs showed low but partial overlapping in carbapenem resistant Pseudomonas aeruginosa genotypes (Golle et al. 2017). Wastewaters from both Austrian private clinics are indirectly discharged into the River Mur where the emergence of ESBL and carbapenemases, dominated by CTX-M ESBL, have been previously described (Zarfel et al. 2017). Wastewaters from the Slovenian hospital are indirectly discharged into the River Drava; both rivers Mur and Drava then flow into the Danube River. Screening for ESBL or carbapenmase harbouring Enterobacteriaceae revealed that 39% of all isolated E. coli and 15% of all Klebsiella spp. from the River Danube had at least one acquired resistance, where the downstream countries (Bulgaria and Romania) have higher resistance rates than countries from the upper regions (Germany, Austria, Hungary). This clearly indicates that the River Danube serves as a reservoir for nearly all clinically important antibiotic resistances in Enterobacteriaceae (Kittinger et al. 2016). Since hospital effluents contribute to less than 2% of the total amount of municipal sewage (Kümmerer 2009b; Carraro et al. 2016), it is plausible that hospitals are not the main source for resistant bacteria in municipal sewage. Therefore, other sources, especially agricultural runoffs and waters used for aquaculture, need to be taken into account and monitored carefully (Lettieri et al. 2018). Different disinfection technologies could be used (Rizzo et al. 2013; Biswal et al. 2014) for eliminating ARB, especially from hospital effluents (Hocquet et al. 2016), but also from WWTP effluents (Jäger et al. 2018; Hembach et al. 2019).

The European Directive 91/271/EEC states that no specific restriction is foreseen for hospital sewage effluents, meaning no pre-treatment of hospital effluents which may be discharged into the main wastewater flow for treatment in WWTP (Hocquet et al. 2016). Therefore, it is necessary to develop additional or adjusted strategies and guidelines adapted for the removal of microbial contaminants in wastewaters, including facultative pathogenic bacteria and ARGs (Jäger et al. 2018). Waters of the Mur, Drava and Danube rivers create a unique area of significant natural and cultural heritage where all three rivers have been nominated for protection under the UNESCO Biosphere Reserve programme. Further research is needed for the protection and sustainable management of transboundary river ecosystems of the Mur, Drava and Danube rivers, also known as ‘the European Amazon’ (UNESCO Biosphere Reserve Mura-Drava-Danube | WWF 2017).

We are very grateful to Sanlas Holding GmbH, Austria and General Hospital Ptuj for cooperation in this project and to Sanlas Holding GmbH, Austria, who acted as a co-financer of the project. The research was financial supported by national research programme (P2-0118).

The Supplementary Material for this paper is available online at https://dx.doi.org/10.2166/washdev.2020.086.