To evaluate the performance of four sampling methods [contact plates, electrostatic wipes (wipe), swabs and a novel roller sampler] for recovery of Staphylococcus aureus from a stainless steel surface.
To evaluate the performance of four sampling methods [contact plates, electrostatic wipes (wipe), swabs and a novel roller sampler] for recovery of Staphylococcus aureus from a stainless steel surface.
Stainless steel test plates were inoculated with Staph. aureus, dried for 24 h and sampled using each of the four methods. Samples were either incubated directly (roller, contact plate) or processed using elution and membrane filtration (swab, wipe). Performance was assessed by calculating the apparent sampling efficiency (ASE), analytical sensitivity (Sn) and percentage of replications with positive growth. The wipe demonstrated the best performance across all inoculating concentrations (ASE48 h = 18%; Sn48 h = 7 CFU per 100 cm2). The swab performed well when corrected for area actually sampled (ASE48 h = 24%; Sn48 h = 76 CFU per 100 cm2). Of the contact-based methods, the newly developed roller sampler outperformed the contact plate (roller: ASE48 h = 10%; Sn48 h = 17 CFU per 100 cm2; contact plate: ASE48 h = 0·04%; Sn48 h = 1412 CFU per 100 cm2); both contact samplers performed better at higher inoculating concentrations (6E3 CFU per 100 cm2 for the roller and 6E6 CFU per 100 cm2 for the contact plate). Overall, the electrostatic wipe produced the highest number of replications resulting in positive growth (74%24 h, 91%48 h).
This study demonstrates that selection of the sampling method must be carefully considered, given that different methods have varying performance.
This is the first study assessing static wipes for sampling and one that uses a more real-world-relevant 24-h drying time. The results help with infection control, and environmental health professionals choose better sampling methodologies.
Staphylococcus aureus is one of the most frequently isolated human pathogens, resulting in significant disease burden (Lowy 1998; Klevens et al. 2007; David and Daum 2010). Methicillin-resistant Staph. aureus (MRSA) is a major source of nosocomial infection and the primary cause of skin and soft tissue infections in the community. In Europe, an extra 5400 deaths per year are attributed to MRSA infections, and in the United States, MRSA infections result in an estimated 19 000 deaths per year – more than AIDS or tuberculosis (ECDC 2009; Klevens et al. 2007; CDC 2011, 2013). Both methicillin-susceptible Staph. aureus (MSSA) and MRSA are capable of asymptomatically colonizing healthy individuals, which is critically important with respect to surface contamination and transmission. In the United States, 28–32% of the general population are colonized with MSSA and 0·8–1·5% with MRSA; colonization rates in Europe vary considerably and are generally lower than in the United States (Kuehnert et al. 2006; Gorwitz et al. 2008). Subgroups, such as healthcare workers, are colonized at even higher rates (Albrich and Harbarth 2008).
Staphylococcus aureus is transmitted primarily through direct contact with a colonized or infected individual. Hands are the critical disseminators, particularly the hands of healthcare workers (Ayliffe et al. 1988; Boyce et al. 1997; Bhalla et al. 2004). However, environmental surfaces and fomites are believed to play an important, but poorly understood, role in transmission (Lowy 1998; Muto et al. 2003; Kazakova et al. 2005; Miller and Diep 2008; Chambers and Deleo 2009). Environmental contamination with Staph. aureus has been reported in numerous studies and is believed to be the source, at least in part, of several hospital outbreaks of MRSA. MRSA contamination has been observed on patient beds, work stations, toilet seats, gowns, staff clothing, stethoscopes and other hospital surfaces (Boyce et al. 1997, 2010; Hardy et al. 2006; Giannini et al. 2009). Staphylococcus aureus is not only easily transmitted from skin to fomites but also, once established, exhibits lengthy survival. Staphylococcus aureus can survive on a variety of surfaces for several months and is highly resistant to desiccation (Dietze et al. 2001; Sexton et al. 2006; Williams and Davis 2009).
Despite the several examples of environmental surfaces and fomites where MSSA/MRSA has been isolated, there are no widely accepted standards for sampling these surfaces. To effectively characterize environmental surface contamination, it is critical to utilize a sampling method that is both sensitive (i.e. the sampler has low detection limits) and efficient (i.e. the sampler captures most of the bacteria present on the surface). Numerous sampling methods (swabs, dipslides, contact plates, electrostatic wipes, etc.) have been employed for routine and postoutbreak environmental surveillance (Sexton et al. 2006; Obee et al. 2007; Cimolai 2008; Felkner et al. 2009; Otter et al. 2009; Sherlock et al. 2009).
An ideal sampling method would satisfy three important criteria. First, the sampler must effectively remove bacteria from the surface being sampled. Second, removal should maximize collection and minimize bacterial damage. Finally, the sampler must release captured bacteria as much as possible for analysis (Davidson et al. 1999; Moore and Griffith 2007). In the case of contact-based samplers, the third requirement does not involve release of the bacteria, but rather the sampler needs to provide optimal conditions for growth because no intermediate processing is performed. Sampling methods should also be chosen in the light of the analytical technique (e.g. culture-based vs PCR-based).
The overarching objective of this study was to evaluate the performance of four sampling methods for the detection of viable Staph. aureus on environmental surfaces. The performance of three currently used sampling methods (contact plates, electrostatic wipes and swabs) and one novel sampling method [a rolling contact-based sampler (roller)] was compared. Using these results, infection control and environmental health investigators will be equipped to make more informed decisions regarding the selection of sampling systems.
A series of experiments on stainless steel plates were conducted to determine the sampling recovery efficiency and analytical sensitivity of a novel roller sampler (in-house design), contact plate (laboratory made, RODAC-style), rayon swab (Fisher Transport Swabs, Waltham, MA, USA) and electrostatic cloth (Swiffer®; Procter and Gamble, Cincinnati, OH, USA). All work was done within a 1300 Series A2 biological safety cabinet (Thermo Scientific, Asheville, NC, USA).
The steel plates were 16-gauge T-304 (no. 4 annealed) stainless steel (40·6 cm × 25·4 cm), demarcated into 100 cm2 zones. Stainless steel was chosen as the experimental test surface because of the material's ability to withstand autoclave sterilization and because numerous hospital and community fomites/surfaces are constructed of this material. All plates were autoclaved prior to inoculation.
Contact-based samplers were prepared using mannitol salt agar (MSA; Becton Dickinson BBL, Sparks, MD, USA), which was chosen as a selective and differential medium that is frequently used as a screening medium for MSSA/MRSA. To prepare the roller sampler (licence pending), a custom autoclavable mould was created; a base insert was placed into the mould, and the mould was filled with molten MSA (Fig. 1). RODAC-style, 60-mm contact plates (PML Biomerieux, Durham, NC, USA) were prepared using appropriate amount of MSA to achieve an appropriate convex meniscus. Regular MSA plates were also made with 47-mm Petri dishes (Millipore, Bedford, MA, USA) for use with membrane filters. All medium plates and rollers were air-dried until set, exposed to ultraviolet light for a minimum of 30 min, covered and refrigerated until sampling.
The stainless steel surfaces were inoculated with a range of concentrations (6E0, 7E1, 5E2, 6E3 and 6E6 CFU per 100 cm2) of MSSA (as a surrogate for MRSA; ATCC 29213, Manassas, VA, USA). Stock suspensions of MSSA were prepared using fresh colonies from blood agar (TSA II; BD BBL) after incubation at 37°C for 24 h. The population density of the stock suspension was adjusted by dilution with saline solution (0·9%) and was set at 0·5 McFarland standard with a nephelometer (TREK, Cleveland, OH, USA). Aliquots of stock suspensions were inoculated onto MSA plates in serial dilution for determining the average starting concentration.
The autoclaved steel plates were visually checked for dryness, and each 100-cm2 zone was inoculated with 100 μl of the MSSA stock suspension (prior to inoculation, a blank control sample was obtained to ensure sterility of the steel). The stock suspension was spread using sterile spreaders for a period of one minute per zone; a new spreader was used for each zone. The steel plates were undisturbed until dry and then placed into a sanitized holding container for 24 h at room temperature.
Samples were collected after 24 h of drying. Each zone was sampled using either the roller sampler, contact plate, swab or static wipe, at each of the five Staphylococcus contamination levels; the sampling order of zones was random. Although the area sampling zones was 100 cm2, only the roller and electrostatic wipe covered the entire zone during sampling; the swab-sampling area was determined to be 36 cm2, and the contact plate covered 28 cm2.
The roller was dried for 10 min at room temperature (inside the biosafety cabinet, to eliminate condensation), placed onto a sampling handle and rolled along the steel surface.
Contact plates were placed against the steel surface for 10 s at a uniform downward pressure using 840 g weights. Weighting of other methods was considered, but was not conducted because in field conditions, weighting is only relevant to the contact plate. In addition, a complex, automated system would be required to equally weight other methods, which would not be repeatable in real-world use.
Rayon swabs (Fisher Transport Swabs) were first premoistened with Amies solution (which was present in the base of the swab-holding tube as a transport medium). Once saturated, the swab was rotated axially and moved laterally in a zigzag motion across the steel surface.
The electrostatic wipe was folded once (in addition to the factory folds) and wiped three times each direction on each side (‘North–South’ and ‘East–West’); the wipe was then turned over, and wiping was repeated (quality control was maintained based on wipe methods described elsewhere (Hoet et al. 2011)). The wipe was then placed into a sterile Whirl-Pak bag (Nasco, Fort Atkinson, WI, USA) for processing. An equal number of passes were made for each method during each replication.
A total of 168 samples were collected [eight replications were conducted for each method at four inoculating concentrations (6E0, 7E1, 5E2 and 6E6), and 10 replications were conducted for each method at one inoculating concentration (6E3)]. Roller samples and contact plates did not require further processing and were placed directly into a 37°C incubator for up to 48 h. Swabs and electrostatic wipes were processed using a membrane filtration procedure for concentrating bacteria (Lutz and Lee 2011). For swab samples, 10 ml PBS–Tween 20 was added to the swab reservoir tube. The swab was vortexed for 10 s to enhance cell removal from the swab head, and the entire suspension was passed through a sterile 0·45-μm membrane filter (Millipore). A similar procedure was employed to process the electrostatic wipes. Fifty millilitres of PBS–Tween 20 was added to each Whirl-Pak bag; the bag was resealed and manually massaged until the PBS solution was absorbed. Each bag was then massaged rigorously for an additional 10 s, and the suspension was passed through a 0·45-μm membrane filter. All membrane filters were handled aseptically and placed directly onto MSA plates and incubated at 37°C for up to 48 h.
All MSA plates and roller samples were counted after 24 and 48 h. To eliminate counting background micro-organisms (which may have been present on the electrostatic wipe, as they were not sterile), only those colonies that were morphologically consistent with Staph. aureus and displayed mannitol fermentation were counted.
The mean concentration [recovered colony-forming units (CFU), normalized to CFU per cm2 sampled] was calculated for each sampling method. Sampling methods were also compared by sampling recovery efficiency, analytical sensitivity and the percentage (frequency) of replications that resulted in positive growth (across all inoculating concentrations). Chi-squared categorical analysis was conducted to evaluate the relationship between growth and sampling method. Statistical analysis was conducted using PASW Statistics 18.0 (SPSS 18.0 SPSS, Chicago, IL, USA); a P-value < 0·05 was considered statistically significant.
Apparent sampling recovery efficiency and analytical sensitivity calculations were based on formulae described by Obee et al. (2007). Apparent sampling recovery efficiency was determined as the ratio of CFU recovered from the test area to the number of CFU inoculated onto the surface (expressed as a percentage). This can be summarized using the formula:
where ASEt, apparent sampling efficiency at incubation time t (t = 24 h or t = 48 h); Ct, concentration of MSSA recovered (in CFU per cm2) at incubation time t = 24 h or t = 48 h; Ci, concentration of MSSA inoculated (in CFU per cm2) at time t = 0 h.
Analytical sensitivity was calculated for 100 cm2 using the formula:
where , analytical sensitivity (CFU required per 100cm2 to observe a positive result) at incubation time t (t = 24 h or t = 48 h); A, area sampled (cm2); ASEt, apparent sampling efficiency at incubation time t (t = 24 h or t = 48 h).
Comparison of the four sampling methods indicated that recovery of Staph. aureus varied widely across methods when tested with various contamination levels. The overall mean CFU recovery was highest for the wipe and the roller. At lower contamination levels (6E0–6E3 CFU per 100 cm2), the wipe performed better than the roller. When corrected for area sampled, swabs demonstrated similar levels of recovery to the wipe and roller. Wipe and swab methods produced higher recovered CFU than contact-based methods at the three lower contamination levels (6E0–5E2 CFU per 100 cm2). In contrast, at the highest contamination scenario (6E6 CFU per 100 cm2), the roller recovered far more CFU than the other methods and, at the second highest contamination level (6E3 CFU per 100 cm2), recovered nearly the same as the wipe.
Apparent sampling efficiency varied considerably across sampling methods, ranging from 0 to 25% at 24 h of incubation (Table 1). The swab method was the most efficient in nearly all cases (mean ASE24 h: 10%, 0–25%), followed by the wipe method (mean ASE24 h: 2%, 0–8%). Again, at the highest inoculating concentration, the roller sampler demonstrated the best performance, resulting in the highest ASE (ASE24 h = 0·01%). Forty-eight-hour incubation markedly improved noted efficiency, with efficiency up to 46% for the swab (mean ASE48 h: 24%) and up to 33% for the wipe (mean ASE48 h: 18%).
|Range of sampling efficiencies||Range of sampling sensitivities||Area sampled (cm2)|
|Mean sampling efficiency at 24-h incubation (%)||Mean sampling efficiency at 48-h incubation (%)||Mean sampling sensitivity at 24-h incubation (per 100 cm2)||Mean sampling sensitivity at 48-h incubation (per 100 cm2)|
|Contact plate||0·01||0·04||10 061||1412||28|
The minimum detection limit, which was expressed as analytical sampling sensitivity (the minimum number of CFU required for a positive result, at a specific inoculation concentration and a given surface area sampled), was determined for each method (Table 1). The wipe and swab methods were far more sensitive than the contact-based methods; that is, they required fewer CFU on the sampling surface to produce a positive result. The analytical sensitivity was as low as 4E0 CFU per 100 cm2 for the swab (mean Sn24 h: 4E3 CFU per 100 cm2) and as low as 1E1 CFU per 100 cm2 (mean Sn24 h: 5E3 CFU per 100 cm2) for the wipe. Among the contact-based methods, the roller sampler was far more sensitive than the contact plate (as low as 6E1 CFU per 100 cm2 for the roller vs as low as 2E3 CFU per 100 cm2 for the contact plate; mean Sn24 h roller: 2E3 CFU per 100 cm2, mean Sn24 h contact plate: 1E4 CFU per 100 cm2); however, the contact plate sensitivity is based upon the highest two concentrations only, as the contact plate did not result in any growth at the lowest three concentrations. Observed analytical sensitivity improved with 48-h incubation, improving to as low as 2E0 CFU per 100 cm2 for the swab (mean Sn48 h: 8E1 CFU per 100 cm2), 3E0 CFU per 100 cm2 for the wipe (mean Sn48 h: 7E0 CFU per 100 cm2), 7E0 CFU per 100 cm2 for the roller (mean Sn48 h: 2E1 CFU per 100 cm2) and 1E3 CFU per 100 cm2 for the contact plate (mean Sn48 h: 1E3 CFU per 100 cm2).
The frequency of replications resulting in positive growth across all contamination levels was determined for each sampling method (Table 2). This dichotomous outcome of ‘growth/no growth’ was expressed as the percentage of times that at least one colony-forming unit was observed among replications of a given sampling method. At 24 h of incubation, the wipe performed the best, and increased incubation time resulted in more frequent positive results, with the wipe resulting in growth 91% of the time. The contact plate showed the fewest positive results. Categorical analysis (based on chi-squared estimates) indicated that there was a significant association between growth and sampling method (P24 = 0·001; P48 < 0·001).
|Mean positive at 24-h incubation (%)||Mean positive at 48-h incubation (%)|
|Chi-squared||P = 0·001||P < 0·001|
This study provides data on the performance of four different sampling methods for the collection of Staph. aureus from environmental surfaces. For the current evaluation, we designed and created a prototype flexible contact-based sampling tool. Due to the novel rolling design of this sampler, it has the potential to overcome two of the major limitations of currently available contact-based sampling methods: lack of flexibility and restriction to sample small surface areas. Subsequently, the newly designed roller was compared with three sampling methods to determine their sampling performance, using a series of controlled experiments. To our knowledge, this is the first study that evaluated electrostatic wipes for Staph. aureus sampling and the first description of a flexible contact-based sampling device. The results indicated that, overall, wipe- and swab-sampling methods were superior to contact-based methods across a range of diverse contamination levels (6E0–6E6 CFU per 100 cm2; using a range of concentrations, as opposed to a single high concentration, the research is more relevant to real-world conditions, where surface concentrations may vary greatly; Table 3). However, at higher contamination levels, the novel roller sampler outperformed all other methods and provided a desirable alternative to standard contact plates.
|Frequency of positive growth (%)a||57||91||57||36|
|Sensitivitya||76 CFU||7 CFU||17 CFU||1412 CFU|
|Low conc. performanceab||Good||Good||Moderate||Poor|
|High conc. performanceab||Moderate||Moderate||Good||Moderate|
|Ability to pool||Not effectively||Yes, effectively||Yes, but effectiveness unknown||Not effectively|
|Ease/time of processing||Moderate||Moderate||Easy||Easy|
|Ability to pre-enrich||Yes||Yes||No||No|
|Field usability||Good||Good, requires strict quality control||Good, but requires additional field study||Good|
|Additional comments||Good for small or delicate surfaces. Poor for large surface areas. May be limited by processing. Vortexing and concentrating or pre-enrichment recommended||Excellent for large or irregular surfaces. May be limited by processing. Vortexing and concentrating or pre-enrichment recommended. Requires strict quality control measures||Better than contact plate/dip slide for large/irregular surfaces. Dirty surfaces result in debris overloading||Good for small, flat surfaces. Recovery may be poor at low surface bioburden. Easy to use with no technical expertise. Dirty surfaces result in debris overloading|
Environmental surface sampling for bacterial contamination is an important tool for routine environmental monitoring, cleaning efficiency evaluations and outbreak investigations. Several sampling devices are available for surface sampling, but few studies have evaluated the performance of these devices for the isolation of Staph. aureus. In their study of several sampling methods for MRSA isolation, Obee et al. (2007) found similar levels of efficiency and analytical sensitivity compared with the current results. Despite the similarities in overall performance, Obee et al. and other researchers have found that contact-based methods are far superior to swabs, contrary to the findings of the current study (Lemmen et al. 2001; Moore and Griffith 2002; Obee et al. 2007).
There are several possible reasons for this discrepancy. First, postinoculation/presampling drying time may significantly affect the degree to which Staph. aureus can be recovered. In previous studies, samples were obtained both immediately after inoculation and 30 min to 1 h after sampling (Moore and Griffith 2002; Obee et al. 2007). A 24-h drying period was used in this study, which may have resulted in poorer recovery. Staphylococcus aureus has been shown to survive well on environmental surfaces for weeks to months; however, die-off may still be considerable after 24 h. Past research has demonstrated that sampling performance is negatively affected even after only 1 h of drying prior to sampling; in some cases, as much as a 5-log difference in recovery (Moore and Griffith 2002; Obee et al. 2007). Although sampling immediately after or within an hour of inoculating the steel surfaces would have likely increased recovery in the current study as well, these short drying periods are less relevant to actual field conditions. Bacteria deposited on common surfaces are unlikely to be removed in a time period <24 h, because cleaning may be infrequent or incomplete. Even in healthcare settings, where surface cleaning is routine, it is doubtful that all common touch surfaces (which may harbour Staph. aureus) are cleaned every day (because cleaning is often focused on aesthetic improvement, rather than on microbial reduction; Dancer 2008). Therefore, it is believed that the bacterial recovery observed here is more likely to represent recovery in real-world settings.
The processing method for swabs and electrostatic wipes may also account for the discrepancy observed between this study and past research. In prior work, swabs were processed using either pour plating or direct swab inoculation (DSI; Moore and Griffith 2002; Obee et al. 2007). Using the pour-plating technique, the swabs were vortexed to enhance bacterial release from the swab; however, pour plating is known to result in lower bacterial counts than spread plating due to contact with hot agar (Massa et al. 1998). Indeed, pour plating resulted in lower recovery in past studies than DSI (Moore and Griffith 2002; Obee et al. 2007). Using DSI, bacterial loss may result from poor release of the bacteria from the swab, because no vortexing is conducted (Lemmen et al. 2001; Moore and Griffith 2002; Obee et al. 2007). Utilizing both a vortexing and concentrating step through membrane filtration, current results indicated good recovery (on a per cm2 sampled level) using the swab and wipe methods.
Incubation time may have also resulted in differences between studies. Obee et al. (2007) conducted plate counting after 24 h of incubation. In the current research, colonies were counted after both 24 and 48 h. The additional 24 h of incubation resulted in much higher observed recovery than at 24 h, demonstrating the benefit of extended incubation times (clearly the number of CFU present is the same at both 24 and 48 h; however, the extended incubation time allows for enhanced colony development and resolution during colony counting).
Despite the findings of this research, the study is not without limitations. The membrane filtration method that was employed to concentrate Staph. aureus is widely used for water quality investigations, but has not been routinely employed for processing surface samples. Because the objective of the study was to evaluate sampling method performance to give maximum recovery, it was felt that the concentrating step may have resulted in higher recovery than without. However, using this new approach may limit comparisons with previous research.
This study affirms the importance of selecting a sampling method appropriate for the given application. Choosing a method solely on the basis of expense, availability or convenience may lead to biased results. Given that the best environmental surface-sampling methods have poor to moderate recovery in a controlled, laboratory setting, using an inappropriate system in real-world applications is likely to result in a higher occurrence of false-negative results. In the case of an outbreak investigation or the study of a high-priority pathogen (such as MRSA), the use of an inefficient method may place additional people at unnecessary risk for infection.
Future research efforts should focus on additional sampling methods and attempt to replicate these results in field conditions.