Population Screening Using Sewage Reveals Pan-Resistant Bacteria in Hospital and Community Samples

The presence of pan-resistant bacteria worldwide possesses a threat to global health. It is difficult to evaluate the extent of carriage of resistant bacteria in the population. Sewage sampling is a possible way to monitor populations. We evaluated the presence of pan-resistant bacteria in Israeli sewage collected from all over Israel, by modifying the pour plate method for heterotrophic plate count technique using commercial selective agar plates. This method enables convenient and fast sewage sampling and detection. We found that sewage in Israel contains multiple pan-resistant bacteria including carbapenemase resistant Enterobacteriacae carrying blaKPC and blaNDM-1, MRSA and VRE. blaKPC carrying Klebsiella pneumonia and Enterobacter cloacae were the most common Enterobacteriacae drug resistant bacteria found in the sewage locations we sampled. Klebsiella pneumonia, Enterobacter spp., Escherichia coli and Citrobacter spp. were the 4 main CRE isolated from Israeli sewage and also from clinical samples in our clinical microbiology laboratory. Hospitals and Community sewage had similar percentage of positive samplings for blaKPC and blaNDM-1. VRE was found to be more abundant in sewage in Israel than MRSA but there were more locations positive for MRSA and VRE bacteria in Hospital sewage than in the Community. Therefore, our upgrade of the pour plate method for heterotrophic plate count technique using commercial selective agar plates can be a useful tool for routine screening and monitoring of the population for pan-resistant bacteria using sewage.


Introduction
Antibiotic-resistant bacteria are a growing problem worldwide and thus possess a major threat to global health. Extensive use of antimicrobial agents in hospital settings and animal husbandry, has led to the emergence of pan-resistant "superbugs", which creates serious ecological and epidemiological challenges for humans [1][2][3][4].
There is increasing concern about the growing resistance of pathogenic bacteria in the environment. These resistant bacteria are excreted by humans and animals and have been found in different environmental compartments. Most typically, in hospital effluent, municipal sewage, surface, ground and drinking water, as well as sediments and soil. [14].
During 2006 Israeli hospitals faced a clonal outbreak of carbapenem-resistant Klebsiella pneumoniae [15]. Since then, several reports have detected CRE's in hospital settings around Israel [16][17][18][19]. As Schwaber et al. noted in their report concerning the spread of bla KPC pathogens in Israel, bla KPC is almost exclusively healthcare acquired with no significant community transmission. Experience with community acquired extended spectrum β-lactamase (ESBL) microorganisms [20][21][22] and the knowledge that long term care facilities (LTCF's) have become a source of reintroduction of CRE to acute care hospitals [23], the spread of CRE to the community is possible and should be examined.
Since there are no standard methods or commercial selective plates for environmental sewage samples which meet our needs, we developed a method that utilizes clinical selective plates for environmental sewage samples analysis.
Here we demonstrate that the detection of pan-resistant bacteria can be done on environmental samples using commercial selective agar plates in a manner that could be used in routine screening of the population. The identity of the microorganism and the drug resistance mechanism can then be identified by classical bacteriological identification, sensitivity tests and Real Time-Polymerase chain reaction (RT-PCR).
Using this method we sought to evaluate the presence of CRE, MRSA and VRE bacteria in Israeli sewage from hospitals and community collected from all over Israel, representing the population carriage of pan-resistant bacteria and to establish a convenient and fast method for sewage sampling.

Sampling Site Information
Two types of sewage sites were analyzed; (1) Sewer systems that drain sewage from the local municipal surroundings and (2) Sewage treatment facility (STF) that drain all the lines of the communities that are included in that line. The sampling locations were divided into four major areas: North, Center, East and South and were categorized as follows: Haifa District: Haifa: Includes the cities Haifa, Nesher, Krayot and Tirat-HaCarmel (STF). Tel Aviv District and Central Region: (1) Tel Aviv: Eight different sampling locations in the city of Tel Aviv-Jaffa (All of the sampling sites in Tel Aviv are sampling sites on sewer systems). (2) ShafDan (Wastewater Treatment Plant of Dan Region): An STF that includes the following cities: Tel Aviv-Jaffa, Ramat-Gan, Bnei-Brak, Giv'atayim, Petah-Tikva, Giv'at-Shmuel, Kiryat-Ono, Tel-HaShomer, Ramat Ef 'al,Azur, Bat-Yam, Holon, Rishon-Lezion and Rehovot. Jerusalem City: Three different sampling sites in the city of Jerusalem, which cover the entire city. Due to the unique mountain topography of Jerusalem, each sampling site stands on its own. Sorek and Og are STF's. South District: Three different locations were sampled: Rahat and Kuseife and the biggest city in the south of Israel Be'er-Sheva. Be'er-Sheva and Rahat are STF's (Fig 1 &  Table 1) (The Israeli Central Virology laboratory is authorized to sample sewage for routine environmental surveillance as part of the ministry of health of the state of Israel). All sampling locations were also divided by the type of sewage drained; Pipelines that drain sewage from hospitals and/or nursing homes (Hospital) were distinguished from those that drain sewage from the community which do not contain hospitals and/or nursing homes (Community).

Collection of sewage samples
All sewage samples were collected using automatic composite samplers [24]. The final samples comprised a mixture of 24 or 48 individual samples gathered over a period of 24 hours. Sampling sites were selected following two criteria; (1) Main entrance to a treatment plant: sampling the enteric microbial content of an entire community's sewer system. (2) Specified sites in the sewer system: sampling the enteric microbial content of specific neighborhoods where there were no hospitals or nursing homes.

Selective overlay plate preparation
Sewage samples were plated according to the pour plate method for heterotrophic plate count technique described in the Standard Methods for the examination of water & waste water [25] with minor modifications. Three different chromogenic media were used as the underlay layer, that are based on the classic chromogenic agar CHROMagar Orientation, enabling easy detection and presumptive identification of bacteria by their specific colored colonies according to their chemical and enzymatic properties; (1) CHROMagarKPC is selective for Gram-Negative Carbapenem-Resistant bacteria, (2) CHROMagarMRSA is selective for Staphyloccocus aureus and (3) CHROMagarVRE is selective for VRE (Hy-Labs, Rehovot, Israel).
Briefly, the overlay agar was prepared by diluting 1:1 one of the chromogenic agars with distilled water. The agar was finely chopped and then gently mixed with slightly cooled down boiled distilled water until melted. 1 ml of sewage was added to 6 ml of the pre-melted and diluted chromogenic agar resulting in an overlay with a final dilution of 1:12. The mixture was gently mixed and then poured into 90 mm plates containing the specific CHROMagar growth media and were incubated at 35°C ± 2 for 48 hours.

Isolation and identification of resistant colonies
From each overlayed CHROMagar plate a certain percentage of suspected to be CRE, MRSA or VRE (according to instructions of manufacturer) were isolated, Gram stained using Aero-sprayWescor, Slide stainer Cytocentrifuge and observed with a light microscope (Olympus BX43, USA). Identification and susceptibility of CRE and MRSA was performed using the Phoenix™ (BD, NJ, USA) and MALDI-TOF MS (microflex LT™, Bruker, Germany) systems. Uncertain bacterial identifications were tested by classical/biochemical methods according to the literature [26,27] and by 16S rRNA sequencing of the first 800bp of the gene (Hy-labs, Israel) for confirmation. Identification and susceptibility of VRE was performed classically according to the literature [26,27].

DNA extraction
Automated Nucleic Acid (NA) extractions were carried out from fresh well-isolated colonies by NuclisensEasyMAG (bioMerieux, France) or by EZ1 advanced (Qiagen, Germany) using

Clinical samples
The clinical samples reviewed in this study were blood cultues, general cultues and rectal swabs for screening received by the Clinical Microbiology Laboratory in The Chaim Sheba Medical Center in Israel during 2013. General cultures include synovial, pericardial, pleural and peritoneal fluids, peritoneal dialysate, brain abscess, lung biopsy/tissue, pericardial fluid/tissue, heart valve, bone, joint, internal organs tissues, biopsies, heart valves, electrodes, catheters and wounds. All cultures suspected to be CRE were tested by Modified Hodge test, MBL E-test and by E-tests for imipenem, meropenem and ertapenem (bioMérieux, France) following CLSI guidelines [30].

Results
Determination of the different assay's limit of detection (LOD) In order to determine the limit of detection (LOD) for the overlay method, 1.5X10 8 CFU/ml bla KPC Klebsiella pneumoniae, bla NDM-1 Escherichia coli, MRSA or VRE positive bacteria were serially dilluted logarithmically in sewage that was negative for these targets. Diluted bacteria were overlaid (as described above) or classically streaked on selective agar. Plates were incubated for 18-24 hours at 35°C ± 2 and analyzed after 48 hours. We found that the overlay did not cause bacteria death as the LOD was similar (5-10 CFU/ml) between the two methods for the different bacteria.  (Table 2).

Selection of the resistant colonies
bla KPC more common in sewage than bla NDM-1 in Israel The presence of pan-resistant bacteria in sewage samples was tested in 16 different sampling locations from April 2012 until November 2013 (Table 1) as described above. Initially we examined the incidence of bla KPC and bla NDM-1 in sewage samples around Israel. bla KPC was detected in most (12/16, 75%) of the locations sampled; Northern Israel, represented by Haifa, had both samplings positive for bla KPC . In the center of Israel which included 8 locations in Tel-Aviv and the ShafDan-STF, 75% of the locations were positive for bla KPC . Jerusalem and the South of the country had 67% locations positive for bla KPC . However, bla NDM-1 was detected only in 2/16 (13%) of the locations sampled; 1 location in Jerusalem and 1 location in the south ( Table 1). The State of Israel was divided into four main areas for sewage sampling; Haifa district in the north, Central district including Tel-Aviv and the Shafdan-STF, Jerusalem City in the East and the South district. Each district was divided to sub-districts. The boundaries and population size are mentioned for each sampling site. The number and percentage of positive samples for pan-resistant bacteria; CRE (bla KPC and bla NDM-1 ), MRSA and VRE out of the number of samplings analyzed is indicated.

The incidence of bla KPC and bla NDM-1 in sewage from Hospitals and the Community
We compared the percentage of positive sewage samples for bla KPC and bla NDM-1 in sewage that contains hospitals and nursing homes (Hospital) to sewage that does not contain hospital and/or nursing homes (Community). We found no significant differences in the percentage of positive samples for bla KPC and bla NDM-1 between Hospital and Community sewage (78% and 63% for bla KPC and 11% and 13% for bla NDM-1 ) (Fig 2).
Among the locations that do contain Hospital sewage, Jerusalem has the lowest percentage of bla KPC positive isolates (19%) compared to the rest of the country (47-71%) but has the most bla NDM-1 positive samples (13%) compared to the rest of the country (0%) ( Table 2).

The distribution of CRE bacteria in Israeli sewage and clinical samples isolates
Using the selective overlay plate technique enabled us to isolate colonies and to study them. Therefore, we isolated 112 CRE colonies from the different sampling locations. Fifty two out of the 112 (46%) CRE colonies isolated around Israel were bla KPC positive by RT-PCR. All bla KPC positive bacteria were further analyzed and the distribution of these isolates by genera is shown in Table 3; Citrobacter braakii (3), Citrobacter freundii (4), Enterobacter asburiae (1), Enterobacter cloacae (14), Escherichia coli (7), Klebsiella oxytoca (4) and Klebsiella pneumoniae (19). Only 3 carbapenem-resistant colonies out of the 112 (3%) were positive for bla NDM-1 by RT-PCR. Final bacteria identification revealed that 2 isolates were of Escherichia coli and 1 was Klebsiella pneumoniae (Table 3).
bla KPC carrying Klebsiella pneumoniae is also the most common CRE isolated from clinical samples in our clinical microbiology lab. The second, third and fourth most isolated CRE bacteria in our lab are E.coli, Enterobacter spp. and Citrobacter spp., respectively. These bacterial species were also the main species found in the sewage (Fig 3B).

VRE is more abundant in sewage in Israel than MRSA
Overall there were many more positive locations for VRE than for MRSA (69% and 19%, respectively) ( Table 1). When examining the differences between Hospital and Community sewage we found more locations positive for MRSA and VRE bacteria in Hospital sewage than in the Community (Table 2). No district had MRSA positive samplings in the Community.
Jerusalem and the South district had no samplings positive for MRSA even in Hospital sewage. VRE was positive at all districts from Hospital and Community sewage except from Community sewage in Jerusalem (Table 2).

Discussion
Since there were no commercial methods for environmental sewage sampling using selective plates to meet our needs, we developed a method that combines the selective ability of the clinical selective plates with the environmental sewage sampling. We found that this method has an average LOD of 5-10 CFU/mL and thus can be easily used by laboratories for sewage screening. We have shown that sewage in Israel contains multiple MDR bacteria including CRE carrying bla KPC and bla NDM-1 , MRSA and VRE.
It is most interesting that while we found in our work that the most abundant CRE in sewage in Israel was bla KPC carrying Klebsiella pneumonia followed by Enterobacter cloacae. (Table 3), in a previous publication by Xinzhuo Zhang et al. bla   positive Citrobacter freundii and Enterobacter cloacae were the most abundant CRE found in hospital sewage in China [31]. The divercification pattern of pan-resistant bacteria in sewage could imply the carrier rate of those organisms in the population. Our finding that bla KPC carrying Klebsiella pneumoniae followed by Enterobacter cloacae and Escherichia coli are the most abundant CRE in Israeli sewage correlates with our additional findings that bla KPC carrying Klebsiella pneumoniae followed by Escherichia coli and Enterobacter spp. are the most common CRE bacteria found in clinical samples as blood and general cultures and in routine rectal swabs screening in our hospital. This finding also correlates with the fact that bla KPC carrying Klebsiella pneumoniae was the cause of an outbreak in Israeli hospitals in 2006 and has become a continuous struggle to control it since [16,17,32]. bla NDM-1 carrying bacteria which have also been reported recently in Israeli hospitals [18,33,34] were also present in Israeli sewage, though to a much lesser extent (13%) than bla KPC (75%) (Tables 1 and 3). Our results show that this method can be a useful tool in screening and maybe predicting bacterial outbreaks in Hospitals and the Community. It is important to emphasize that, though the community is not supposed to be exposed to sewage, once in a while there is an intentional or not flowing of sewage to the ocean or leakage of sewage pipelines which exposes the community to unwanted microorganisms including pan-resistant bacteria. Therefore, we cannot be indifferent towards the microorganism content of sewage as pan-resistant bacteria such as CRE may cause serious infections, especially bacteremia and urosepsis and carries a high associated mortality rate [35][36][37][38][39]. Our study also shows that generally Hospital and Community sewage contain similar types of pan-resistant bacteria in Israel (Fig 2 & Table 2). This is surprising as we would expect to find less resistant bacteria in the community. Therefore, it is not clear if the presence of pan-resistant bacteria in the community is due to person to person transmission having hospitals a significant source of these bacteria or due to other environmental factors such as water or food consumption such as intensively farmed chicken carrying VRE [40] but it emphasizes that those bacteria are prevalent in the community.
Our study shows many more locations positive for VRE than MRSA in Israeli sewage (69% and 19%, respectively (P = 0.0078) (Fig 2) but data published by the Israeli National Center for Infection Control (2013 annual summary-general hospitals) shows much higher incidences of MRSA bacteremia in Israeli general hospitals than VRE (%20% and 5%, respectively /100,000 patients a day). The increased presence of VRE upon MRSA in Hospital and Community sewage may be due to the differences in the bacteria's preferences. MRSA is typically nasal or  dermal while VRE is intestinal. There could also be differences in survival rates of the bacteria in sewage environments. Though sewage is an environment, containing water and nutrients, it may have extreme living conditions such as fluctuating levels of energy sources, oxygen, temperature, pH and osmolarity. [41]. Many polluting factors such as metals, chemicals, detergents and oils resulting from the industry may also affect bacteria survival in sewage. We conclude that (1) our upgrade of the pour plate method for heterotrophic plate count technique using commercial selective agar plates can be a useful tool for routine screening and monitoring of the population for pan-resistant bacteria using sewage and (2) that the content of MDR bacteria in sewage represents the distribution of pan-resistant bacteria in the population and that (3) similar resistant bacteria are present in the Hospitals and in in the Community in Israel.
More research should be conducted in order to estimate the possible use of environment sampling to predict outbreaks of pan-resistant bacteria in the population and to look for other resistance genes such as bla OXA-48 , bla VIM and bla IMI in sewage systems.

Author Contributions
Conceptualization: LMG YM SGH MH EM NK.