JSTOR

Methods

Environmental samples were obtained in a 9‐bed general ICU, which is an open unit with no side rooms. During an initial 6‐month period (April through September 2002), environmental screenings were performed monthly, followed by 8 environmental screenings performed over the course of 6 months (July through December 2003) and 10 more environmental screenings performed during January and February 2004, to investigate closely the changes in environmental contamination with MRSA. Throughout the study, cleaning in the ICU remained unchanged and complied with National Health Service standards.⁸ Detergent and water were used for cleaning all surfaces, with duties being split between a domestic worker and a housekeeper, both of whom worked during the morning. All environmental screenings took place after the completion of cleaning at 2:00 p.m. At each environmental screening, samples were taken from the same 4 areas in all 9 bed spaces: underneath the bed, the workstation, the control buttons on the monitors, and a ledge positioned behind the bed. At each site, an area of approximately 10 cm² in size was swabbed by rotating a sterile cotton swab, which had first been immersed in phosphate‐buffered saline, in 3 directions. The swab was immediately placed in 10 mL of brain‐heart infusion broth. After incubation at 37°C overnight, the bacterial suspension was subcultured onto oxacillin‐resistant screening agar base and a Baird Parker plate agar (Oxoid Unipath). Presumptive positive colonies were confirmed as S. aureus, and resistance to methicillin was determined using a multiplex polymerase chain reaction for the mecA and coa genes.⁹ In addition to environmental screening, samples were obtained from patients' nose, perineum, and wound sites at admission and then 3 times weekly. Both broth enrichment in brain‐heart infusion broth and direct inoculation were used for detection of MRSA in samples from patients. No isolation facilities were available for patients from whom MRSA was isolated; therefore, barrier nursing care was performed. All environmental MRSA isolates and multiple patient MRSA isolates were typed using pulsed‐field gel electrophoresis (PFGE), according to a standardized method using SmaI restriction endonuclease (Invitrogen), as described elsewhere.¹⁰ Multiple patient isolates were selected for typing from different sites and time points throughout the patients' stay in the ICU. PFGE results were entered into BioNumerics (Applied Maths) and were analyzed using the unweighted pair group method using arithmetic averages. Profiles with a difference of more than 1 band were considered to be distinct. Statistical analysis was performed using Pearson's correlation coefficient, to determine the relationship between patient colonization and environmental contamination.

Results

Typing of Environmental and Patient Isolates

A total of 250 environmental MRSA isolates, including individual isolates from 188 environmental sites and multiple isolates from 31 sites, were typed by PFGE. Thirty‐eight patients colonized with MRSA were present during the environmental screenings, and 139 isolates from those patients were typed by PFGE. Multiple isolates were typed from all patients (range, 2‐22 isolates), apart from 3 patients for whom only 1 isolate was detected.

Sixteen different PFGE profiles were identified from the environment, of which 13 (D, E, F, H, I, J, L, M, N, O, P, X, and Y) had a high degree of similarity (more than 80%) to the UK strain EMRSA‐15 (Figure 1). Three PFGE profiles (B, T, and W) had a high degree of similarity to strain EMRSA‐16. Eleven different PFGE profiles were identified from the patients, with 9 having a high degree of similarity to UK EMRSA‐15, of which 7 (D, I, J, L, N, O, and P) were also isolated from the environment. The remaining 2 profiles (B and T), both of which were isolated from the environment, had a high degree of similarity to the UK strain EMRSA‐16. Ten PFGE profiles occurred at more than 1 environmental screening; PFGE profile J, which was isolated at 14 (58.3%) of the 24 screenings, was the most predominant. The other most frequently occurring PFGE profiles were I, O, and P, each of which was isolated at 11 (45.8%) of 24 screenings. Despite being the most predominant strain in the environment, PFGE profile J was not the most predominant among patients, with more patients being colonized by strains with PFGE profiles B, I, O, and P. At 20 of the environmental screenings, more than 1 PFGE profile was isolated from the environment.

(74 KB)

Figure 1.  Pulsed‐field gel electrophoresis (PFGE) profiles of methicillin‐resistant Staphylococcus aureus isolated from the environment, the number of environmental screenings during which each of the different PFGE profiles was identified, and the correlation with PFGE profiles of strains colonizing patients.

Open New Window

When the PFGE profiles of the strains isolated from the patients were compared with those isolated from their immediate environment (ie, underneath the bed, the monitors, and the workstations within their bed space), indistinguishable strains were isolated on 20 (35.7%) of 56 occasions (Figure 2). When the area of sampling was extended to establish whether the colonizing strains were present anywhere in the environment in the ICU at that point in time, it was found that 32 (57.1%) of 56 patients were colonized with strains that were indistinguishable from the environmental strains. Ten PFGE profiles were present in the environment at times when no patients colonized with strains with the same PFGE profiles were present.

(50 KB)

Figure 2.  Correlation between the predominant pulsed‐field gel electrophoresis (PFGE) profiles of methicillin‐resistant Staphylococcus aureus strains isolated from the environment and patients during 24 screenings performed over the course of 2 years.

Open New Window

Acquisition of MRSA by Patients in the ICU

During the second 8‐month period of environmental screening, 61 patients in the ICU were colonized with MRSA; 35 of the patients had been colonized before admission, whereas 26 of the patients acquired MRSA while staying in the ICU. Of these 26 patients, 12 of them acquired a strain that was indistinguishable from a strain with which at least 1 other patient in the ICU was colonized during their stay. However, 14 patients acquired a strain that was distinct from the strains colonizing other patients during their stay.

When the extent of environmental contamination and patient acquisition are correlated, differences between the PFGE profiles emerge. For PFGE profile B (EMRSA‐16 subtype), both importation and acquisition by patients and environmental contamination occurred during a particular, defined period. Colonization of patients by the other 4 major MRSA strains (I, J, O, and P) was spread over the 8‐month period, but the spread of environmental contamination was different. PFGE profiles I and P were not represented in environmental screenings conducted during the first 6 months, despite the presence of colonized patients in the ICU, whereas the environmental contamination with PFGE profiles J and O was much more extensive.

Three of the 14 patients who acquired MRSA in the ICU at times when no other patients colonized with MRSA of the same type were present acquired it within 10 days of the same type being isolated from the environment (Figure 3). No patients with indistinguishable strains had been present in the ICU for at least 7 days previously; in the case of patient 411, no patients with an indistinguishable strain had been present in the ICU for 54 days previously.

(33 KB)

Figure 3.  Evidence for transmission of methicillin‐resistant Staphylococcus aureus (MRSA) with pulsed‐field gel electrophoresis (PFGE) profiles J, O, and P from the environment to patients in an intensive care unit. The time lines highlight the dates when MRSA was isolated from the environment, when the patients were admitted, and when they acquired MRSA. During the time that these patients were hospitalized, no other patients in the intensive care unit were colonized with MRSA with the same PFGE profile.

Open New Window

Discussion

Widespread contamination of the environment with MRSA was found at times when both small and large numbers of patients colonized with MRSA were present, possibly indicating the existence of a secondary reservoir of MRSA, in addition to the patients themselves. The site underneath the beds had the highest levels of MRSA contamination, possibly because it is not far from the floor, which has been cited in previous studies as having the highest levels of contamination in settings where MRSA is endemic.⁶ Although the floors are infrequently touched by hands, floors, especially those directly under beds, may play a role in the transmission of MRSA by transferring the MRSA via the movement of dust in air currents to surfaces that are touched more frequently. Shiomori et al.¹¹ demonstrated that MRSA carried on dust particles was capable of being aerosolized and, indeed, was present in the respirable range. During sampling, the workstation and the monitors were observed to be touched by staff much more frequently than the area underneath the bed; the workstations were used to prepare drugs and store other disposable items that were used frequently, and the monitors were continually being adjusted. Therefore, although the workstations and monitors have lower levels of contamination than do the area under the beds, the consequences of contamination of the workstations and monitors are potentially greater, in terms of transmission of MRSA to patients.

Although the percentage of sites from which MRSA was isolated was higher when a patient colonized with MRSA was present in the bed space (25.4%), MRSA was isolated from 20.2% of sites, even when there was not a patient colonized with MRSA in the bed space. This finding raises several issues in terms of infection control policies. Currently, the guidelines recommend increased terminal cleaning after discharge of MRSA‐colonized patients, which would, therefore, not include the bed spaces that had environmental contamination but did not have MRSA‐colonized patients, potentially leaving bed spaces contaminated with MRSA.¹² Other questions raised are the origin of contamination of these sites: did they become contaminated by previous patients within the bed spaces or by patients in other bed spaces? Our study shows that, when colonized patients are in the ward, MRSA strains from patients may be found in the environment some distance from the colonized patient.

Although variants of EMRSA‐15 predominated, the number of different PFGE profiles isolated from the environment in the present study was far greater than that in other studies, in which one strain has predominated. The higher number of strains identified probably reflects the endemic presence of MRSA situation in our ICU, compared with the presence of MRSA during an outbreak, the setting in which most previous studies have been conducted.¹³ The range of PFGE profiles isolated from the environment reflects the PFGE profiles colonizing patients. However, on only 35.7% of occasions were strains indistinguishable from those isolated from another patient in the colonized patient's immediate environment. This situation is unlike the situation reported previously by Boyce et al.,⁶ where the patient and environmental isolates from the room in which the patient was receiving nursing care were indistinguishable. Unlike the present study, in which the ICU was an open ward, all of the patients in the study by Boyce et al.⁶ were in side rooms; therefore, the spread of the environmental MRSA was more contained. In an early study of the spread of S. aureus in an open ward, results similar to those of our study were observed, with the same strain being isolated from the bedding and from the occupant of the bed on only 27 of 72 occasions.¹⁴ However, as in the present study, the environmental strains were more often identical to those colonizing another patient elsewhere in the ward, possibly indicating spread of the bacteria.

It is known that MRSA has the ability to survive within the environment,^5,15 and, although it was not possible in the present study to determine how long a strain had survived within the environment, many of the strains were found in the environment after a patient with an indistinguishable strain had been discharged, possibly indicating that the strain had survived in the environment. Interestingly, those strains that did not survive in the environment for a prolonged period were present at only 1 or 2 environmental screenings. Differences in the survival of different MRSA strains have been described elsewhere,¹⁶ and there is some evidence from the present study to support this finding. The environmental survival of PFGE types J and O appears to differ, despite both strains being present in high numbers of patients. PFGE type O was not present in the environment if patients colonized with PFGE profile O were not present, with the converse being true with strain J.

Transmission by the hands of HCWs, with colonized patients acting as reservoirs, is thought to be the main route of transmission of MRSA, but more than one‐half of the patients who acquired MRSA in the ICU acquired a strain that was distinct from any strain colonizing other patients during their stays. From this finding it can be hypothesized that a secondary MRSA reservoir exists that is leading to transmission to patients. There is evidence suggesting that at least 3 patients (239, 319, and 411) acquired MRSA from the environment while in the ICU. All 3 patients were admitted to the ICU shortly after the isolation of MRSA with an indistinguishable PFGE profile from the environment, and no patients with indistinguishable strains were present in the ICU for at least the preceding 7 days. Despite the evidence pointing to the environment playing a role in MRSA transmission, in the cases of these 3 patients, some caution in this interpretation should be exercised, because other explanations are possible. For example, colonized HCWs or MRSA‐colonized patients who had been admitted for <48 h but who were not identified by the study may have transmitted the strain. The study did not include the screening of HCWs, and, therefore, we can not conclude that transmission was not from colonized HCWs; however, in previous studies in which HCWs have been colonized with MRSA, colonization has been connected with outbreaks of a specific MRSA strain. This phenomenon was not observed in the present study, and, therefore, transmission by a colonized HCW seems to be a less likely possibility.

In many studies of MRSA environmental contamination in side rooms, the strain colonizing the patient is indistinguishable from that contaminating the environment. In many countries, it is not possible for patients to be receive nursing care in side rooms, and our study is the first to investigate the environment on an open ward over a considerable period by use of detailed typing. It reveals that it is a very complex situation and confirms the importance of physical separation of patients and, if this is not possible, the need for more intensive and effective environmental cleaning measures to eliminate MRSA in the ICU.