Antibiotic resistant Staphylococcus aureus in hospital wastewaters and sewage treatment plants with special reference to methicillin-resistant Staphylococcus aureus (MRSA)

Authors

  • J.M. Thompson,

    1. Faculty of Science, Health and Education, University of the Sunshine Coast, Maroochydore, QLD, Australia
    Search for more papers by this author
  • A. Gündoğdu,

    1. Faculty of Science, Health and Education, University of the Sunshine Coast, Maroochydore, QLD, Australia
    Search for more papers by this author
  • H.M. Stratton,

    1. School of Biomolecular and Physical Sciences, Griffith University, Nathan Campus, Brisbane, QLD, Australia
    2. Smart Water Research Centre, Southport, QLD, Australia
    Search for more papers by this author
  • M. Katouli

    Corresponding author
    1. School of Biomolecular and Physical Sciences, Griffith University, Nathan Campus, Brisbane, QLD, Australia
    2. Smart Water Research Centre, Southport, QLD, Australia
    • Faculty of Science, Health and Education, University of the Sunshine Coast, Maroochydore, QLD, Australia
    Search for more papers by this author

Correspondence

Mohammad Katouli, School of Health and Sport Sciences, University of the Sunshine Coast, Maroochydore DC 4558, QLD, Australia. E-mail: mkatouli@usc.edu.au

Abstract

Aims

To investigate the presence of methicillin-resistant Staphylococcus aureus (MRSA) in untreated hospital wastewaters (UHWW), their transmission into the receiving sewage treatment plant (STP) and survival through the STP treatment.

Methods and Results

Over eight consecutive weeks of sampling, we isolated 224 Staph. aureus strains from UHWW-1, UHWW-2 and its receiving STP inlet (SI) and post-treatment outlet (SO). These strains were typed using the PhP typing method and RAPD-PCR and tested for their antibiotic resistance patterns. Resistance to cefoxitin and the presence of mecA gene identified MRSA isolates. In all, 11 common (C) and 156 single (S) PhP-RAPD types were found among isolates, with two multidrug resistant (MDR) C-types found in H2, SI and SO. These C-type strains also showed resistance to cefoxitin and vancomycin. The mean number of antibiotics to which the strains from UHWW were resistant (5·14 ± 2) was significantly higher than the STP isolates (2·9 ± 1·9) (P < 0·0001). Among the 131 (68%) MRSA strains, 24 were also vancomycin resistant. MDR strains (including MRSA) were more prevalent in hospital wastewaters than in the STP.

Conclusion

This study provides evidence of the survival of MRSA strains in UHWWs and their transit to the STP and then through to the final treated effluent and chlorination stage.

Significance and Impact of the Study

This preliminary study identifies the need to further investigate the load of MRSA in hospitals' wastewaters and possible their survival in STPs. From a public health point of view, this potential route of hospital MRSA dissemination is of great importance.

Introduction

Antibiotic resistant bacteria (ARB) are present and increasing in many countries worldwide causing great concern for public health (Islam 2011). Excessive use of antimicrobial agents to treat human infectious disease in both clinical settings and the general community, as well as in intensive livestock industries, has led to the increased prevalence of ARB (Wise et al. 1998; Mazel and Davies 1999; Islam 2011). Antibiotics place a high selective pressure on bacteria, thus eliminating the antibiotic susceptible strains and subsequently causing the proliferation and dissemination of resistant bacteria. Resistance genes can also be transferred between cells on plasmids or transposons by transductive or conjugative processes (Berger-Bachi 2002). DNA elements that mediate integration of resistance genes (e.g. integrons) may also be involved (Moura et al. 2012), resulting in the further spread of MDR bacteria.

There are two principal categories of ARB of concern for public health. Community-acquired (CA) strains are generally found in nonclinical settings and can affect otherwise healthy individuals. They are generally acquired via close person-to-person contact and affect mainly skin and soft tissue. Hospital-acquired (HA) strains are found in clinical settings, occur in people with predisposing risk factors and more often lead to invasive disease (Miller et al. 2005). HA strains are of particular concern as treatment options are limited and expensive (Tiemersma et al. 2004; Magiorakos et al. 2012). One particularly notorious HA bacterium is the opportunistic pathogen Staph. aureus. Although part of the normal microflora of skin (Nimmo et al. 2006), it can cause a wide variety of conditions in hospitalized patients ranging from minor skin infections (Kummerer 2004; McCaig et al. 2006) to scalded skin syndrome, impetigo, toxic shock syndrome, pneumonia and postoperative infections with the risk of septicaemia (Miller et al. 2005; Moran et al. 2006; Tacconelli 2009; Turnidge et al. 2009). Methicillin-resistant Staph. aureus (MRSA) strains are of high importance as they show resistance to all β-lactam antibiotics because of the presence of the mecA gene. This gene codes for the production of a penicillin-binding protein (PBP) (Ubukata et al. 1989), which renders a number of antibiotics ineffective.

It is known that Staph. aureus and MRSA may gain a temporary residence within hospitals (Chambers and DeLeo 2009), therefore posing the possibility of dissemination into the environment through routes such as hospital wastewater (Baquero et al. 2008). It is proposed that these bacteria may then travel to sewage treatment plants (STPs) through the sewerage system, possibly survive the treatment process and find their way into the environment via STP effluent (Baquero et al. 2008). It has been shown that MRSA strains could also be multiresistant (Farzana et al. 2011); therefore, the release of these hospital strains could contribute to the gene pool for MDR bacteria in the treatment plants and the environment. One study by Borjesson et al. (2010) identified and characterized MRSA at different preliminary treatment stages within a wastewater treatment plant; however, these workers did not identify MRSA from sewage outlet post-treatment. Therefore, while the prevalence and pathogenesis of HA MRSA has been widely studied (Wise et al. 1998; Chang et al. 2003; Nimmo et al. 2006), information on their survival in the STP and the environment once they are released from hospitals is minimal.

We hypothesized that MDR Staph. aureus strains (including MRSA) present in hospitals may be discharged via wastewater and survive during their transit to the receiving STP. We proposed that owing to their prevalence and persistence in hospitals (Wise et al. 1998; European Centre for Disease Prevention and Control 2009), MDR Staph. aureus strains could be frequently found in hospital wastewater, and despite the high dilution rate occurring in the sewer system, they can still be found in the receiving STP if an intensive sampling is performed. In view of the aforementioned one, this study was undertaken to determine the prevalence of antibiotic resistant Staph. aureus strains and MRSA in hospital wastewater, their transport to the receiving STP and survival through the STP treatment stages.

Materials and methods

Sites and sampling

One regional hospital (H1) was originally investigated to identify the prevalence of antibiotic resistant bacteria (ARB) in its wastewater against that of a metropolitan hospital (H2). With H2, we further extended our investigation to detect the survival of ARB during their transmission to the receiving STP by sampling from the STP inlet and outlet after biological treatment and chlorination. Both hospitals had one main wastewater outlet pipe for the entire hospital. From H1, weekly samples were collected from the main outlet for 12 weeks at 7·00 am. From H2 and its receiving STP, weekly samples were collected for 8 weeks at 10·30 am and at 11·00 am of the same day, respectively. The sewer channel taking hospital wastewater to the STP was estimated to be 12 km. The STP was an activated sludge plant with N and P reduction and services an equivalent population of 130 000 and has a 12- to 13-day sludge age. Samples were collected from the incoming raw sewage and treated effluent after the activated sludge treatment and chlorination. The final effluent is discharged to a nearby waterway. Samples were processed in accordance with the Australian and New Zealand Standards for Water Microbiology and Water Quality Sampling (Australian and New Zealand Standards 1998a,b, 2007). In brief, wastewaters were collected in 500-ml sterile microbiological containers mounted onto a handle of appropriate length using ‘grab-sampling’ technique. They were transported to the laboratory on ice and processed within 4 h of collection.

Isolation and identification of Staphylococcus aureus and MRSA

Hospital wastewater and STP incoming influent (SI) samples were processed using serial dilutions, and the STP outgoing effluent (SO) was processed by membrane filtration. Direct and filtered samples were cultured on Vogel-Johnson agar (Oxoid), but samples from hospital 2 and its receiving STP were additionally cultured on mannitol salt agar (Oxoid) containing 6 μg ml−1 of cefoxitin, as a surrogate for methicillin (Broekema et al. 2009). After 24- to 48-h incubation at 37°C, suspected Staph. aureus colonies that showed a positive catalase reaction were transferred to nutrient broth (Oxoid) containing 20% glycerol and stored at −80°C for further analysis.

DNA extraction and polymerase chain reaction (PCR)

Chromosomal DNA was extracted using genomic DNA extraction mini kit (Bioline, Alexandria, NSW, Australia) following the protocol for Gram-positive bacteria with minor adaptations. In brief, cells were harvested from an overnight nutrient broth (Oxoid, Adelaide, SA, Australia), 4 ml of the suspension was centrifuged for 10 min at 4500 g, and the pellet was resuspended in 180 μl of enzymatic lysis buffer for 30 min at 37°C. Cells were then treated with 25 μl of proteinase K and 200 μl of buffer AL and incubated for a further 30 min at 56°C. After treating with 200 μl of ethanol (96–100%), samples were washed twice with buffer and centrifuged for 1 min at 11 700 g. DNA was eluted from the column using 200 μl of elution buffer, incubated at room temperature for 1 min and centrifuged at 5200 g for another minute. Purified DNA was stored at −20°C for PCR.

PCR was performed using Staph. aureus species-specific primers for the nucA gene, which codes for thermostable nuclease (Pinto et al. 2005) to confirm the species on a genetic level rather than solely on biochemical results. The primer sequences used were F 5′-GCGATTGATGGTGATACGGTT-3′ and R 5′-AGCCAAGCCTTGACGAACTAAAGC-3′ (Pinto et al. 2005) generating a 270-base pair fragment. PCR amplification was performed according to the study by Pinto et al. (2005) using a reaction mixture containing a master mix of 30·9 μl of filter-sterilized Milli-Q water, 5 μl of 10× PCR buffer (Bioline), 0·5 μl dNTP (10 mmol l−1) (Fisher Biotech, Wembley, WA, Australia), 1·5 μl MgCl2 (50 mmol l−1) (Bioline), 5 μl of forward and reverse nucA primers (10 μmol l−1) (Invitrogen, Mulgrave, VIC, Australia), 0·1 μl Taq polymerase (5 U μl−1) (Bioline) and 2 μl of purified DNA. The PCR protocol was denaturation for 5 min at 94°C; 35 cycles of 30 s at 94°C, 45 s at 62°C and 45 s at 72°C; and a final extension step of 10 min at 72°C. Amplified PCR product was electrophoresed on a 2% agarose gel in a 0·6× Tris base–EDTA (TBE) buffer and stained with ethidium bromide and viewed under UV light.

Antibiotic resistance testing

In accordance with Clinical Laboratory Standard Institute (CLSI) standards (CLSI 2011), all Staph. aureus strains isolated from hospital wastewaters, SI and SO were tested for the their resistance against eight antimicrobial agents using the following antimicrobial impregnated discs (Oxoid): tetracycline (30 μg), amoxicillin-clavulanic acid (20/10 μg), ampicillin (10 μg), gentamicin (10 μg), ciprofloxacin (5 μg), chloramphenicol (30 μg), amikacin (30 μg), cefoxitin (30 μg) (a surrogate for methicillin) (Munckhof et al. 2009). All isolates were also tested for their resistance against vancomycin using the agar dilution method recommended by CLSI standards (15) with 8 and 32 μg ml−1 vancomycin. For isolates that exhibited resistance against cefoxitin and vancomycin (32 μg ml−1), minimum inhibitory concentration (MIC) was determined using the E-test strips (Oxoid M.I.C.E. test) (Brown and Brown 1991) for the antibiotics oxacillin (representing methicillin resistance) and vancomycin as per manufacturer instructions.

PCR confirmation of methicillin resistance gene

In view of the discrepancies found in the literature regarding the presence or absence of the mecA gene in MRSA isolates (Suzuki et al. 1992; Bignardi et al. 1996), all Staph. aureus strains were tested for the presence of the mecA gene. The primer sequences used were F 5′-CCTAGTAAAGCTCCGGAA-3′ and R 5′-CTAGTCCATTCGGTCCA-3′ (Yadgar et al. 2009) that generate a 314-base pair fragment. PCR amplification was performed as described by Yadgar et al. (2009) using a reaction mixture containing a master mix of 10·75 μl of filter-sterilized Milli-Q water, 2·5 μl of 10× PCR buffer (Bioline), 0·25 μl dNTP (10 mmol l−1) (Fisher Biotech), 1·5 μl MgCl2 (50 mmol l−1) (Bioline), 0·25 μl of forward and reverse mecA primers (10 μmol l−1) (Invitrogen), 0·2 μl Taq polymerase (5 U μl−1) (Bioline) and 2 μl of purified DNA. The PCR cycle consisted of denaturation for 5 min at 95°C; 35 cycles of 2 min at 95°C, 1 min at 58°C and 1 min at 72°C; and a final extension step of 10 min at 72°C. Amplified PCR product was electrophoresed, and bands were visualized as described previously. All PCRs included one sample lacking a template as a negative control and an in-house positive control MRSA strain (8 M14) proved to have mecA and nucA gene after sequencing.

Typing of isolates

All isolates were typed using a high-resolution biochemical fingerprinting method specifically developed for typing of staphylococci (PhP-CS plate; PhPlate Microplate Techniques AB, Sattsjo-Boo, Sweden) and RAPD-PCR. For biochemical fingerprinting, a loopful of a fresh bacterial culture was inoculated in 10 ml of PhPlate growth media containing 0·2% (w/v) proteose peptone, 0·05% (w/v) yeast extract, and 0·5% (w/v) NaCl and 0·011% (w/v) bromothymol blue. Aliquots of 175 μl of each bacterial suspension were inoculated into the 24 wells of each set by the aid of a multichannel pipette. The plates were then incubated at 37°C, and images of the plates were scanned after 16, 40 and 64 h using a HP Scanjet 4890 scanner (Hewlett-Packard, Palo Alto, CA, USA). After the final scan, the PhPlate software (PhPWin ver. 4.2; PhP Microplate Techniques AB) was used to create absorbance data (biochemical fingerprint) from the scanned images according to the manufacturer's instructions. Similarity among the isolates was calculated using the correlation coefficient after a pairwise comparison of the biochemical fingerprints and clustered according to the unweighted pair group method (UPGMA) with arithmetic averages (Sneath and Sokal 1973). Isolates having the same fingerprint were regarded as belonging to the same PhP type.

RAPD-PCR was performed as previously described using the random sequence primer S 5′-TCACGATGCA-3′ (Naffa et al. 2006). RAPD-PCR mixture contained the following: 38·5 μl of filter-sterilized Milli-Q water, 5 μl of 10× PCR buffer (Bioline), 1 μl dNTP (10 mmol l−1) (Fisher Biotech), 1·5 μl MgCl2 (50 mmol l−1) (Bioline), 2·5 μl of primer (10 μmol l−1) (Invitrogen), 0·5 μl Taq polymerase (5 U μl−1) (Bioline) and 1 μl of purified DNA. The PCR cycle consisted of denaturation for 5 min at 95°C; 35 cycles of 1 min at 95°C, 1 min at 37°C and 2 min at 72°C; and a final extension step of 5 min at 72°C. RAPD-PCR bands were scored with data coded as a factor of 1 or 0, representing presence or absence of each band. Using the PhP software (ver. 4.2), the banding patterns obtained were compared pairwise and clustered as described previously. Isolates belonging to the same PhP-RAPD type were considered as members of the same clonal group and classified as common (C) types.

Statistical analysis

Differences in the prevalence of phenotypically positive MRSA and the strains carrying mecA genes as well as the number of MRSA-positive strains in different samples were assessed using Fisher's exact test. P-value of <0·05 was considered statistically significant.

Results

Common and persistent MRSA in hospital wastewater and STP

Overall, 224 Staph. aureus strains were isolated from H1 (n = 57), H2 (n = 85) and its receiving STP inlet (SI) (n = 74) and outlet (SO) (n = 8). Typing of these isolates showed the presence of 11 common (C) (n = 68) and 156 single (S) PhP-RAPD types altogether (Table 1). The 57 strains from H1 belonged to 1 C-type and 50 S-types, whereas the 85 isolates from H2 and its receiving STP consisted of 10 C-types and 106 S-types. Strains belonging to the 11 C-types were shown to be either present in more than one sample from hospital wastewater but never found in STP samples (e.g. C1) or present in both the hospital wastewaters and the STPs (e.g. C3). Among the 10 C-types from H2, two (i.e. C7 and C8) were found in the STP inlet as well as outlet and were regarded as persistent C-types. We also found 3 C-types (e.g. C2) among the isolates from the STP inlet that were found on more than one occasion (Table 1). All isolates belonging to the 10 C-types from H2 and STP (61 isolates) were mecA positive, but only 32 (53%) of them were phenotypically resistant to cefoxitin. From H1, all isolates from C11 were cefoxitin resistant, but only two were mecA positive (Table 1). Of all C1–C10 isolates, nine isolates were also vancomycin resistant Staphylococcus aureus (VRSA) and were found in H2 (n = 3), SI (n = 5) and SO (n = 1) (Table 1). No C11 isolates were VRSA.

Table 1. Strains of Staphylococcus aureus and methicillin-resistant Staph. aureus belonging to common types (C1–C11) found over 8 weeks of sampling from hospital 1 (H1), hospital 2 (H2) wastewater and its receiving sewage treatment plant (STP). Number of strains that were resistant to vancomycin and/or cefoxitin is also given. Brackets indicate STP inlet isolates. Brackets and underline indicate STP outlet isolates
Common Type (no. of isolates)SourceWeeks where the strains were found
W1W2W3W4W5W6W7W8    
  1. W1–12, weeks 1–12; SI, STP influent; SO, STP outlet; VRSA, vancomycin resistant Staphylococcus aureus; VISA, vancomycin intermediate Staphylococcus aureus.

  2. a

    n number of VISA/VRSA isolates.

  3. b

    n number of cefoxitin-resistant isolates.

  4. c

    n number of mecA-positive isolates (all isolates from C1–10 are mecA positive).

C1 (13)H22b22b28b801b1000    
C2 (2)SI10000001    
C3 (9)H2 (SI)0(3)2b0(1b1)3b300    
C4 (17)H2 (SI)004a1b4 (13a5)00000    
C5 (2)H2 (SI)000001b1 (1)00    
C6 (2)SI001b101000    
C7 (6)H2 (SI) (SO)000(1)1(2)(2)0    
C8 (5)H2 (SI) (SO)00001 (1a1b1)(2)10    
C9 (3)H2 (SI)000002a2b2 (1)00    
C10 (2)SI00001b1b00    
  W1W2W3W4W5W6W7W8W9W10W11W12
C11 (7)H1000000001b1c1006b5c1

Isolates belonging to S-types (n = 156) constituted 70% of the total number of isolates tested. Thirty-five of 50 S-types (70%) from H1 were cefoxitin positive, and only 13 (37%) carried the mecA gene. For H2, 46 of 57 (81%) strains showed resistance to cefoxitin. However, there was a significant difference (P < 0·0001) between the numbers of mecA-positive strains in wastewater of these two hospitals, where 93% of the strains carried this gene (Table 2). The number of mecA-positive strains in wastewater from H2 was also significantly (P < 0·0001) higher than those found in the STP inlet. Twenty-two of the 156 MRSA-positive strains were also VRSA, of which all strains were mecA positive (Table 2).

Table 2. Prevalence of methicillin-resistant Staphylococcus aureus (MRSA) and vancomycin resistant Staphylococcus aureus/vancomycin intermediate Staphylococcus aureus (VRSA/VISA) strains belonging to single PhP-RAPD types found in hospital 1 (H1), hospital 2 (H2) and its receiving STP inlet (SI) or outlet (SO). MRSA strains were identified based on their resistance to cefoxitin
SourceNo. of IsolatesNo. of cefoxitin resistant strains (mecA positive)No. of VRSA/VISA strains
  1. a

    P < 0·0001 for mecA-positive strains in H2 vs H1.

  2. b

    P < 0·0001 for mecA-positive strains in H2 vs SI.

H15035 (13a)12
H25746 (53a)5
SI4516 (39b)2
SO42 (4)1
Total15699 (109)20

Antibiotic resistance patterns of Staphylococcus aureus from hospital wastewater and STP

The pattern of antibiotic resistance among the Staph. aureus strains varied in samples collected from different sites. In all, 192 (86%) of the 224 Staph. aureus strains tested were resistant to between two and nine antibiotics of which 131 (68%) strains were MRSA as shown by their phenotypic resistance against cefoxitin with 24 being also VRSA (Table 3).

Table 3. Antimicrobial resistance patterns for Staphylococcus aureus (including MRSA strains) isolated from hospital 1 (H1), hospital 2 (H2) and its receiving STP
Resistance Source
Total No. of isolatesNo. of isolates
H1H2SISO
  1. TE, tetracycline; AMC, amoxicillin/clavulanic acid; AMP, ampicillin; CN, gentamicin; CIP, ciprofloxacin; CHL, chloramphenicol; AK, amikacin; FOX, cefoxitin; VAN, vancomycin; MRSA, methicillin-resistant Staphylococcus aureus; SI, STP influent; SO, STP outlet.

  2. MRSA strains (FOX) have been underlined, and VRSA (VAN) strains have been boxed.

None60060
Single antimicrobial2627161
 AMP2327140
 TE10010
 CHL10001
 CN10010
Two antimicrobials2747115
 AMP/AMC111271
 AMP/FOX30300
 AMP/AK10010
 AMP/TE60132
 AMP/CHL10100
 AMP/CIP20002
 CHL/AK11000
 CHL/CN22000
Three antimicrobials208480
 AMP/AMC/FOX72140
 AMP/TE/CHL30210
 AMP/TE/CIP10100
 AMP/TE/CN10010
 AMP/AK/FOX10010
 AMP/TE/FOX21010
 AMP/CIP/AK22000
 AMP/AK/AMC22000
 CN/CIP/AK11000
Four antimicrobials (including AMP)2942221
 AMC/CIP/FOX30210
 TE/AMC/AK10010
 TE/CIP/FOX10010
 TE/CIP/CHL10010
 TE/AMC/FOX20020
 AK/AMC/FOX20020
 TE/CHL/FOX20020
 TE/AMC/CHL1000100
 FOX/AMC/VAN10001
 AMC/AK/VAN11000
 CN/FOX/CIP11000
 AMC/CIP/AK11000
 CHL/CIP/AK11000
 TE/AK/CIP20020
Five antimicrobials (including AMP)42102750
 TE/AMC/CIP/FOX80800
 TE/AMC/CHL/FOX20020
 TE/AK/CN/FOX10100
 TE/AMC/CN/FOX1431100
 AMC/CIP/CN/FOX61500
 AMC/AK/CN/FOX20200
 TE/AMC/CHL/VAN30030
 CHL/AMC/CN/AK11000
 CHL/CN/CIP/VAN11000
 CN/FOX/CIP/VAN11000
 AMC/CN/CIP/AK11000
 CHL/CN/FOX/AK11000
 CHL/AMC/CN/FOX11000
Six antimicrobials (including AMP)40102910
 TE/AMC/CIP/FOX/VAN20200
 TE/AMC/CIP/FOX/CN31110
 AMC/CIP/CHL/FOX/CN2131800
 AMC/AK/CIP/FOX/CN80800
 TE/CHL/CN/CIP/AK11000
 CHL/AMC/CN/CIP/AK11000
 AMC/CN/FOX/CIP/AK11000
 TE/CHL/AMC/CN/FOX11000
 TE/CHL/AMC/CN/CIP11000
 TE/CHL/AMC/CN/AK11000
Seven antimicrobials (including AMP)209650
 TE/AK/AMC/CIP/CHL/FOX50230
 TE/AMC/CIP/FOX/CN/VAN10100
 TE/AK/AMC/CIP/FOX/CN30300
 TE/AK/AMC/CHL/FOX/CN21010
 TE/AK/AMC/CHL/FOX/VAN10010
 TE/CHL/AMC/CN/FOX/CIP44000
 CHL/AMC/CN/FOX/CIP/AK22000
 TE/AMC/CN/FOX/AK/VAN22000
Eight antimicrobials (including AMP)42200
 TE/AK/AMC/CIP/FOX/CN/VAN10100
 TE/AK/AMC/CIP/CHL/FOX/CN32100
Nine antimicrobials (including AMP)108101
 TE/AK/AMC/CIP/CHL/FOX/CN/VAN108101

Samples collected from the STP inlet contained the highest number of isolates (n = 22) susceptible to all or resistant to only one antibiotic (32%) compared with those collected from H1 (4%), H2 (8%) and SO (13%). In contrast, Staph. aureus strains isolated from H2 wastewater showed resistance to five or more antibiotics (n = 65; 76%), which was significantly higher (P < 0·0001) than that found in SI (n = 11; 15%) and SO which had only one isolate (Table 3). The mean number of antibiotics to which the strains from the hospitals were resistant to (5·14 ± 2) was significantly higher than STP isolates (2·9 ± 1·9) (P < 0·0001). In all, 93 cefoxitin-resistant strains were found to be present in either H2 wastewaters or the STP (SI and SO) or both, of which two were also found in STP outlet samples (Table 3).

Within the strains isolated from wastewaters from both hospitals, the highest resistance was observed towards ampicillin (93% for H1 and 100% for H2) followed by amoxicillin/clavulanic acid (72% for H1 and 82% for H2), gentamicin (62% for H1 and 76% for H2) and cefoxitin (59% for H1 and 78% for H2) (Fig. 1a). A similar pattern of resistance was observed when only MRSA-positive strains were considered (Fig. 1b). Strains from wastewater from H2 consisted of the highest number of MRSA as indicated by their resistance to cefoxitin (78%) compared with those from H1 (59%) and SI (31%) (Fig. 1a).

Figure 1.

Prevalence of antibiotic resistant Staphylococcus aureus (a) and MRSA (b) strains (indicated by their phenotypic resistance to cefoxitin) isolated from wastewaters of hospital 1 (H1) and hospital 2 (H2) and its receiving sewage treatment plant (STP) inlet (SI). Owing to low number of isolates in STP outlet (Staph. aureus n = 8, MRSA n = 2), prevalence was not included. TE, tetracycline; AMC, amoxicillin/clavulanic acid; AMP, ampicillin; CN, gentamicin; CIP, ciprofloxacin; CHL, chloramphenicol; AK, amikacin; FOX, cefoxitin; VAN, vancomycin; MRSA, methicillin-resistant Staph. aureus. (■) H1; (image_n/jam12037-gra-0001.png) H2 and (image_n/jam12037-gra-0002.png) S1.

The MIC of 50 representative MRSA strains belonging to common and single types found in wastewater of H2 and STP (SI and SO) was determined using the E-test. The results from H2 wastewater showed that for oxacillin 87% had a MIC of >256 μg ml−1, with 10% showing MIC of 16 μg ml−1 and only 3% had a MIC of 0·25 μg ml−1. In contrast, all isolated from STP (SI and SO) had a MIC > 256 μg ml−1 (data not shown). Only 10% of VRSA strains from H2 had a MIC of >256 μg ml−1, and the remaining isolates had a MIC < 8 μg ml−1 with the most common MIC being 4 μg ml−1, found in 53% of isolates. Similar MICs were found among the VRSA from STP where 6% of the isolates had a MIC > 256 μg ml−1 and the remaining isolates had a MIC below 8 μg ml−1 with 67% having an MIC of 4 μg ml−1 towards vancomycin.

Discussion

To our knowledge, this is the first study that investigated the movement and survival of Staph. aureus from hospital wastewater to STP and its presence in the discharged effluent. Typing of the isolates using a combination of a high-resolution PhP typing and RAPD-PCR confirmed that certain clonal groups of Staph. aureus were commonly found in wastewaters from both hospitals. These strains were regarded as common types (i.e. C1–C11). The total number of strains belonging to C-types was higher than the number of Staph. aureus belonging to single (S) types. However, the prevalence and thus the diversity of S-types were always high in both the hospital wastewaters and the STP samples. We postulated that the C-types that constituted 30% of the isolates were resident strains in hospitals, therefore were likely to be consistently found in hospital wastewaters.

To trace the movement of these strains through the sewage collection system to the STP, we initially measured the flow of the wastewater in H2; based on the length of the sewer system (approx. 12 km) to the receiving STP, we calculated a transitional period of 12·5 h for the hospital wastewater to reach its receiving STP. We also postulated that there would be a high dilution of wastewater and thus its bacterial contents, during its transmission through the sewer system. On the basis of these factors as well as the logistical problem of collecting STP samples 12·5 h after initial hospital sampling, we decided to extend our sampling number for 8 weeks to increase the chance of detecting resident strains from the STP. Using this sampling protocol, we were able to isolate some of these hospital clones from the inlet of the STP and showed that they belonged to the same PhP-RAPD types. Interestingly, most of these strains had an identical or very similar antibiotic resistance pattern. Other studies have either identified Staph. aureus in hospital wastewater (Nunez and Moretton 2007; Ekhaise and Omavwoya 2008) in biosolids from an STP (Burtscher and Wuertz 2003) or in some cases failed to identify at all (Rusin et al. 2003), but have not investigated the movement or survival of these strains to the extent done in our study.

Originally, we were interested in identifying the presence and survival of MRSA strains in both the hospital wastewaters and the STP, but in view of the high level of antibiotic resistant Staph. aureus strains in hospital wastewaters, and in view of the high diversity of the strains belonging to S-types, we made a comparison between the level of antibiotic resistance among the strains from hospital wastewaters and a STP and found that hospital strains were significantly more resistant to antibiotics compared with STP strains. Furthermore, the prevalence of MRSA strains in hospital wastewaters was also much higher than that in STP, which reflects similar results that have been reported by others (Schwartz et al. 2003; Volkmann et al. 2004). This could be due to high selective pressure in the hospital as a result of high usage of antibiotics, as it has been demonstrated that hospital wastewater contains a concentration of antibiotics as much as 100 times higher than STPs (Kummerer 2004; Baquero et al. 2008).

In our study, we confirmed the presence of MRSA strains in all samples both by their resistance to cefoxitin and by the presence of the mecA gene. Volkmann et al. (2004) reported the presence of MRSA strains in hospital wastewater only based on the identification of mecA gene and showed that this gene was not common in the municipal wastewater samples. A similar study has been carried out by Borjesson and co-workers (Borjesson et al. 2009) who successfully identified the mecA gene and Staph. aureus in different stages of an STP; however, these researchers did not identify culturable Staph. aureus in the end product after sewage treatment, nor did they address survival of certain MRSA clones through the STP process. However, the aforementioned workers suggested that geographical location might favour the presence and survival of Staph. aureus and MRSA in the outlet of the STP (Borjesson et al. 2009). In our case, Australia has a higher prevalence of MRSA infection in hospitals (Australian Institute of Health and Welfare 2011) compared with Sweden where the study was conducted and where there are vigilant infection control mechanisms that have caused MRSA to remain at low levels (Pettersson et al. 2010). Our study was also based on eight consecutive weekly sampling events to increase the chance of detection of these bacteria, whereas the study by Borjesson et al. (2009) was a seasonal study based on monthly sampling for 1 year.

The original aim of sewage treatment was not for specific pathogen control; however, we know that STPs do significantly reduce pathogen loads in wastewater (Spellman 1999). Of the different groups of resistant Staph. aureus strains (including those resistant to MRSA and VRSA) found in the hospital wastewaters and the incoming samples of the STP, only eight resistant strains were recovered from STP outlet samples. This finding suggests that there was a notable reduction in the number of resistant strains of Staph. aureus during their transport to the receiving STP and throughout STP processes, thus proving treatment to be quite effective in this pathogen's removal. However, some strains may genetically have the ability to survive environmental conditions as suggested for other bacteria (Walk et al. 2007). In our study, three of the eight resistant strains of Staph. aureus that survived the STP treatment were MRSA and/or VRSA.

In this study, we identified 24 different type strains of VRSA among the isolates tested (i.e. 224) according to the CLSI agar dilution method. These strains distributed among three common types and 11 antibiotic resistance patterns. However, when tested using E-test strips, only three of them had a MIC of >256 μg ml−1 (data not shown) with the remaining isolates showing a MIC of <8 μg ml−1. Therefore, these isolates should probably be regarded as vancomycin intermediate Staph. aureus (VISA) according to the CLSI methods, as VRSA is classified as isolates that are resistant to >8 μg ml−1 (CLSI. 2011).

One interesting finding in our study was that the use of cefoxitin as an indicator for methicillin resistance (Broekema et al. 2009) did not always correspond with the presence of the mecA gene. A number of isolates from H2 and STP harboured this gene but did not show resistance to cefoxitin in vitro. Conversely, among the H1 isolates, a number of strains showed phenotypic resistance against cefoxitin without carrying the mecA gene. This absence of mecA with isolates exhibiting resistance has been reported before (Bignardi et al. 1996). In some cases, it has also been discovered that despite showing methicillin-resistant gene presence, there is no production of penicillin-binding protein 2a (PBP2a), the protein that is coded for by mecA gene (Suzuki et al. 1992). This was, however, in coagulase-negative Staphylococcus species, not Staph. aureus. The difference in phenotypic resistance and the presence of mecA can also be attributed to the femA gene. This chromosomally encoded factor involved in the pentaglycine side chain formation of peptidoglycan is essential for the expression of methicillin resistance, and any mutation in that gene affects expression of resistance (Hurlimann-dalel et al. 1992). There is also reason to believe that there are yet to be identified determinants of transcriptional control associated with the mecA gene, which may affect expression of PBP2a. Therefore, despite the presence of the mecA gene, the bacterium may not be able to produce the PBP2a necessary for phenotypic resistance because of transcriptional problems with the gene (Oliveira and de Lencastre 2011). The discrepancy in gene presence and resistance may also lie in the methods used to identify methicillin resistance. It has been suggested by CLSI standards that cefoxitin is an appropriate substitute for indentifying methicillin resistance (Broekema et al. 2009; CLSI. 2011). However, it is also possible that this antibiotic may not always be effective in detecting methicillin resistance in Staph. aureus compared with other antibiotics such as oxacillin.

In our study, the numbers of MRSA strains were higher in H2 than H1 despite that fact that the latter hospital was sampled for four extra weeks. This might partially be due to using only one selective agar without antibiotics during sampling from this hospital. The use of cefoxitin supplementing the mannitol salt agar during sampling from H2 wastewaters was shown to have an impact on our rate of isolation of MRSA strains. It is also possible that the time of sampling (7·00 am for H1 and 10·30 am for H2) might have had an effect on the rate of isolation of these strains as more activities, including patients showering and visits of medical staff, may occur mid-morning compared with early morning. Hospital two was also a metropolitan hospital, compared with H1 being a regional hospital. Therefore, the size of the hospitals could also be a contributing factor in regard to the number of MRSA present in the wastewater.

In conclusion, we found that there were many C-type MRSA strains in hospital wastewater that were able to survive and were transported to the inlet of the STP. However, only three isolates originating in H2 were also found in the treated effluent released to the environment via the sewer outlet. One of these strains was resistant to all nine antibiotics tested. From the data presented here, it would appear that hospitals are adding to the load of MDR Staph. aureus entering STPs. The significance of this for public health is not clear. Further work must be undertaken to characterize and quantify the input of MDR Staph. aureus from hospitals compared with those originating in the general community or other wastewater-related sources. More hospitals and corresponding STPs should also be investigated to determine whether these results are observed elsewhere or isolated to this particular hospital and STP investigated. Only then can it be judged whether hospital wastewater should be subject to pretreatment before release into the sewerage system.

Acknowledgements

This research was undertaken and funded as part of the Urban Water Security Research Alliance, as scientific collaboration in Southeast Queensland, Australia, between the Queensland Government, CSIRO, the University of Queensland and Griffith University and the University of the Sunshine Coast. We acknowledge the assistance and cooperation of Queensland Health, the staff at the two hospitals included in the study and the local municipality involved. The logistics of sampling from the hospitals was complicated and required the cooperation of hospital staff, researchers and the local government agency. For confidentiality, the hospitals and local government organization cannot be named.

Ancillary