Journal of Hospital Infection (2006) 63, 47e54

www.elsevierhealth.com/journals/jhin

Reduction in MRSA environmental contaminationwith a portable HEPA-filtration unit

T.C. Boswell*, P.C. Fox

Department of Microbiology, Nottingham City Hospital, UK

Received 27 May 2005; accepted 22 November 2005Available online 3 March 2006

KEYWORDSMRSA; Settle plates;HEPA filtration;Environmentalcontamination

Summary There is renewed interest in the hospital environment as a po-tentially important factor for cross-infection with methicillin-resistantStaphylococcus aureus (MRSA) and other nosocomial pathogens. The aimof this study was to evaluate the effectiveness of a portable high-efficiencyparticulate air (HEPA)-filtration unit (IQAir Cleanroom H13, Incen AG, Gold-ach, Switzerland) at reducing MRSA environmental surface contaminationwithin a clinical setting. The MRSA contamination rate on horizontalsurfaces was assessed with agar settle plates in ward side-rooms of threepatients who were heavy MRSA dispersers. Contamination rates weremeasured at different air filtration rates (60e235 m3/h) and comparedwith no air filtration using Poisson regression. Without air filtration,between 80% and 100% of settle plates were positive for MRSA, with themean number of MRSA colony-forming units (cfu)/10-h exposure/plateranging from 4.1 to 27.7. Air filtration at a rate of 140 m3/h (one patient)and 235 m3/h (two patients), resulted in a highly significant decrease incontamination rates compared with no air filtration (adjusted rate ratios0.037, 0.099 and 0.248, respectively; P< 0.001 for each). A strong associ-ation was demonstrated between the rate of air filtration and the meannumber of MRSA cfu/10-h exposure/plate (P for trend< 0.001). In conclu-sion, this portable HEPA-filtration unit can significantly reduce MRSAenvironmental contamination within patient isolation rooms, andthis may prove to be a useful addition to existing MRSA infection controlmeasures.ª 2005 The Hospital Infection Society. Published by Elsevier Ltd. All rightsreserved.

* Corresponding author. Address: Department of Microbiology, Nottingham City Hospital NHS Trust, Hucknall Road, NottinghamNG5 1PB, UK. Tel.: þ44 115 9691169x45572; fax: þ44 115 9627766.

E-mail address: tboswell@ncht.trent.nhs.uk

0195-6701/$ - see front matter ª 2005 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved.doi:10.1016/j.jhin.2005.11.011

48 T.C. Boswell, P.C. Fox

Introduction

Handwashing by healthcare workers is thought to bethe most important method of control of methicillin-resistant Staphylococcus aureus (MRSA) and othernosocomial infections.1 However, there is renewedinterest in the hospital environment as a potentiallyimportant factor in cross-infection.2e4 The authorsrecently investigated an outbreak of Hickman-line-associated MRSA bacteraemia in a haematologyday-case unit. As part of the investigations, signifi-cant MRSA contamination of the environment wasfound, and settle plates were positive for MRSA ona number of occasions, suggesting ongoing dispersalof MRSA into the air. At least two patients were iden-tified as heavy shedders of MRSA, and the outbreakwas controlled when a single room was designatedfor their further day-case treatment.

This outbreak prompted an investigation of tech-nologies that may reduce future MRSA environmen-tal contamination within such a setting. A number ofair decontamination products have been reviewedrecently by the Rapid Review Panel of the HealthProtection Agency, including air filtration units andultra-violet air disinfection units (www.hpa.org.uk/infections/topics_az/rapid_review). Two of theseproducts have been recommended for in-use evalua-tions in a clinical setting (recommendation level 2),including the IQAir Cleanroom H13 portable airpurification unit (Incen AG, Goldach, Switzerland).This article presents an evaluation of the effective-ness of the IQAir unit at preventing contaminationof environmental surfaces with MRSA in three differ-ent clinical settings.

Methods

IQAir Cleanroom unit

Two IQAir Cleanroom H13 units were kindly providedby Air Science Ltd (Stroud, UK). The machines werefree-standing floor units operating in recirculationmode. This portable air filtration purifier removesairborne micro-organisms by repeatedly drawingambient room air through a pre-filter and thenthrough a high-efficiency particulate air (HEPA)filter, with an efficiency of 99.97% for particlesgreater than or equal to 0.3 mm. The IQAir H13 unit’scentrifugal fan can be set to run at five speeds,with air filtration rates ranging between 60 and400 m3/h.

Patients and clinical setting

The IQAir unit was evaluated in the environment ofthree patients who were known to have MRSA

colonization and who were likely to be heavydispersers of MRSA.

Patient A was a 61-year-old male with acutemyeloid leukaemia. He had previously receiveda bone marrow transplant complicated by graft-vs-host disease of his skin, and subsequent relapse ofhis leukaemia. MRSA colonization was first detectedin April 2004, and MRSA was repeatedly detectedfrom subsequent nose, perineal and skin swabsdespite several courses of Aquasept washes andBactroban nasal cream. This patient regularly at-tended the haematology day-case unit for bloodproduct support, and had been previously linkedwith the MRSA outbreak on that unit. Since then, hehad continued to receive day-case treatment (ap-proximately once every two weeks) in a designatedside-room on an elderly assessment ward, directlyadjacent to the haematology day-case unit.

Patient B was a 66-year-old male who was aninpatient on a medical gastroenterology ward fol-lowing complicated upper gastrointestinal surgeryand a prolonged stay in the intensive care unit. MRSAcolonization was first detected in March 2005, andMRSA was subsequently identified from nose, peri-neum, sputum (from tracheostomy), wound swabsand abdominal drain sites despite two courses ofAquasept/Bactroban. The patient was receivingtotal parenteral nutrition via a central line, re-quired regular suctioning of his airway, and haddeveloped a number of pressure sores. Although ina side-room on the ward, the door had often beenleft open and this patient had been linked witha cluster of MRSA infections on the ward.

Patient C was a previously fit and healthy42-year-old male admitted with severe Stevens-Johnson syndrome following a course of penicil-lin. This had resulted in extensive skin blistering,akin to a severe burn injury. Six days afteradmission, MRSA was detected in sputum, bloodcultures and, subsequently, numerous woundswabs. The patient was mechanically ventilated,hyperpyrexial (requiring haemofiltration for tem-perature control), and was having a completedressing change every 48 h. MRSA sepsis wastreated with intravenous vancomycin and genta-micin, but topical MRSA treatment was notcontemplated.

Study design

Quantitative assessment of MRSA environmentalsurface contamination was measured with agarsettle plates. Environmental surface contamina-tion rates were measured both with and withoutair filtration on separate days (see Table I).

Portable HEPA filtration and MRSA 49

With Patient A, settle plates were used on sixoccasions over an eight-week period and the use ofthe IQAir unit at a filtration rate of 235 m3/h (N¼ 3)was alternated with no air filtration (N¼ 3). WithPatient B, settle plates were used on eight occa-sions over a two-week period, first with no air filtra-tion (N¼ 2) and then with the IQAir unit running atair filtration rates of 140 m3/h (N¼ 2), 95 m3/h(N¼ 2) and 60 m3/h (N¼ 2). With Patient C, settleplates were used on four occasions over a one-week period, first with no air filtration (N¼ 2) andthen with the IQAir unit running at a filtrationrate of 235 m3/h (N¼ 2). The air volume of therooms and corresponding air filtration rates interms of air changes/h for each study day areshown in Table I. The layout of each of the roomsand the position of the IQAir units within themare shown in Figure 1.

None of these patients received topical MRSAtreatment during the study periods, and Patients Aand B did not receive any systemic antibiotictreatment. Patient C was treated with intravenousvancomycin and gentamicin throughout the studyperiod.

Settle plates

Standard-sized circular plates containing oxacillinresistance screening agar base (ORSAB) (Oxoid Ltd,Basingstoke, UK) were used as settle plates.Between four and 34 settle plates were placed indesignated positions in the three rooms and ex-posed to the air for between 1.5 and 8 h (see TableI). Within each room, identical plate positionswere used on each study day where possible. Thesettle plate positions used in each room are shownin Figure 1. Peripheral positions on the floor weremost common, but settle plates were also placedunder the bed, on window ledges, at the side ofsinks, on overbed tables and on other horizontalsurfaces where possible. Settle plate MRSA colonycounts were expressed as colony-forming units(cfu)/10 h of exposure in order to calculatea standardized rate of MRSA environmental surfacecontamination over time.

Bacterial isolates and identificationof MRSA

ORSABplateswere incubatedfor48 hat37�C,andthenumber of typical dark-blue colonies of MRSA wascounted. Representative colonies from each platewere confirmed as S. aureus by agglutination for pro-tein A and clumping factor (Slidex Staph Plus, bioMer-ieux, Basingstoke, UK), and for DNAse production.

Methicillin resistance was determined after 24 h ofincubation at 30�C on Columbia agar base supple-mented with 2% oxacillin (PO0879A, Oxoid Ltd).

Statistical analysis

Statistical analysis was performed using EGRETsoftware. Poisson regression was used to calculateadjusted rate ratios to compare the number ofMRSA cfu/10-h exposure/plate, with and withoutthe IQAir unit (controlled for the study day). Poissonregression was also used to look at the effect ofdifferent machine speeds on MRSA colony countscompared with no air filtration, and to calculate a Pvalue for trend through the rate ratios.

Results

Effect of air filtration on MRSAenvironmental contamination

Without air filtration, all three patients wereconfirmed to be heavy dispersers of MRSA (seeTable I and Figure 1). Overall, between 80% and100% of the settle plates were positive, with themean number of MRSA cfu/10-h exposure/plateranging from 4.1 to 27.7. The extent and distanceof MRSA shedding without air filtration for each pa-tient is shown graphically in Figure 1.

Air filtration at a rate of 235 m3/h (Patients Aand C) and 140 m3/h (Patient B) resulted in a sig-nificant decrease in MRSA surface contaminationin each case (Table I). With Patient A, the adjustedrate ratio comparing the mean number of MRSAcfu/10-h exposure/plate with and without air fil-tration was 0.099 [95% confidence intervals (CI)0.077e0.128]; a reduction of approximately 90%.Similar adjusted rate ratios for Patients B and Cwere 0.037 (95% CI 0.017e0.081) and 0.248 (95%CI 0.215e0.285); reductions of 96% and 75%,respectively.

The effect of these air filtration rates on thepattern and distance of MRSA shedding for eachpatient is also shown in Figure 1.

Effect of different air filtration rateson MRSA environmental surfacecontamination

Air filtration at three different rates (60, 95 and140 m3/h) was used with Patient B. There wasa highly significant association between the reduc-tion in MRSA surface contamination and the vol-ume of air filtered (see Table I). Rate ratios

Table I ccus aureus (MRSA) environmental contamination

Patient(room ai

re)

Total no. ofMRSA cfu

Mean MRSAcfu/10-h

exposure/plate

Adjusted rate ratio(95% CI) P value

Patientsingle ro(27.7 m3

10 16.7 e319 12.981 4.9

4 4.0 0.099 (0.077, 0.128)P< 0.00110 0.5

25 1.1

Patientward, si(39.2 m3

74 5.6 e58 4.1

61a 5.1 0.639 (0.505, 0.808)8 0.6

18 1.1 0.233 (0.166, 0.326)10 1.0

4 0.3 0.037 (0.017, 0.081)P (for trend)< 0.0010 0.0

Patientcare, sin(46.2 m3

330 23.6 e233 27.7

74 5.3 0.248 (0.215, 0.285)P< 0.001105 7.5

cfu, coloa 37 cfub Patien

50T.C

.Bosw

ell,

P.C.

Fox

Effect of portable high-efficiency particulate air filtration on methicillin-resistant Staphyloco

and settingr volume)

Date IQAir filtrationrate (m3/h)

IQAir filtrationrate

(air changes/h)

No. ofsettle

plates used

No. (%) ofMRSA þve

plates

Exposutime (h

A, day-case,om)

01.12.04 0 0 4 4 (100) 1.517.12.04 0 0 34 34 (100) 7.2502.02.05 0 0 30 26 (87) 5.6

01.12.04 235 8.5 4 2 (50) 228.01.05 235 8.5 34 7 (21) 6.517.02.05 235 8.5 34 17 (50) 7

B, medicalngle room)

21.04.05 0 0 20 16 (80) 7.7522.04.05 0 0 20 18 (90) 7

03.05.05 60 1.5 20 16 (80) 604.05.05 60 1.5 20 8 (40) 6.5

27.04.05 95 2.4 20 8 (40) 828.04.05 95 2.4 20 7 (35) 5.25

25.04.05 140 3.6 20 4 (20) 626.04.05 140 3.6 20 0 (0) 6.5

C, intensivegle room)

05.05.05 0 0 20 20 (100) 706.05.05b 0 0 12 12 (100) 7

11.05.05b 235 5.1 20 14 (70) 713.05.05b 235 5.1 20 13 (65) 7

ny-forming units.on one agar settle plate, 0e3 cfu on remainder.t had complete change of dressings whilst settle plates exposed.

Portable HEPA filtration and MRSA 51

Figure 1 (aec) Room schematics, position of settle plates and position of IQAir machine (A) are shown for PatientsAeC, respectively. The results of the settle plates on the paired study days for each patient with and without IQAirfiltration are combined. These are summarized as follows: open circle, settle plate with no methicillin-resistantStaphylococcus aureus (MRSA); grey circle, settle plate with 0.1e5.0 MRSA colony-forming units (cfu)/10-h exposure;black circle, settle plate with >5.0 MRSA cfu/10-h exposure. cm h, cubic metres per hour.

comparing the mean number of MRSA cfu/10-h ex-posure/plate with and without air filtration were0.639 (95% CI 0.505e0.808) at a filtration rate of60 m3/h, 0.233 (95% CI 0.166e0.326) at 95 m3/h,and 0.037 (95% CI 0.017e0.081) at 140 m3/h(P value for trend< 0.001).

Discussion

It is well established that MRSA contamination ofthe environment occurs in the vicinity of MRSApatients, and this is the principal reason forisolating patients in single rooms.5 However, rela-tively few reported studies have used agar settle(sedimentation) plates to assess the rate of MRSAcontamination of environmental surfaces.6 Thissimple methodology was used in the present studyto identify heavy shedders of MRSA and to measurethe effect of using a portable HEPA-filtration uniton the contamination rates of environmental sur-faces. The main reason for selecting this method-ology instead of volumetric air sampling was thatthe authors were less concerned with determining

the rate of MRSA shedding into the air, and moreinterested in whether recirculating HEPA filtrationcould reduce the rate of MRSA settling on horizon-tal surfaces within the rooms of known MRSA dis-persers. Whilst settle plates may be regarded asa relatively crude measure of airborne contamina-tion, they do provide a simple and cost-effectiveway of enumerating the contamination rate ofhorizontal surfaces at multiple points within anoccupied side-room. Contact plates or otherenvironmental sampling methods were not usedbecause the horizontal environmental surfaceswithin the occupied rooms could have been con-taminated with MRSA at the start of each studyday, dependent on the time and effectiveness ofthe previous room cleaning.

Despite attempts to keep the settle plateexposure times constant, there was inevitablysome variation in the duration of exposure of thesettle plates on different study days (for practical,logistical and sometimes clinical reasons), andrelatively long exposure times were used (up to8 h). The rate of recovery of MRSA may have beenaffected by some degree of dehydration of the

52 T.C. Boswell, P.C. Fox

agar, and the results may have underestimated thetrue amount of MRSA dispersal. However, it seemsunlikely that these variations could account for thesignificant differences observed in contaminationrates with and without air filtration.

Patients who are likely to be heavy dispersers ofMRSA include those with skin conditions or woundsheavily colonized by MRSA. These patients areoften responsible for the spread of infection.5,7

All three of the study patients had MRSA-colonizedskin conditions and two had been associated withcross-infection to other patients. In addition, onepatient had a tracheostomy and MRSA in the spu-tum which, in the authors’ experience, is a signifi-cant risk for cross-infection. The highest rate ofenvironmental contamination occurred from thepatient with Stevens-Johnson syndrome. It waseducational for the infection control team andthe healthcare workers on the busy intensivecare unit to see the number of MRSA cfu on the set-tle plates (330 cfu on 20 settle plates equates tow2750 cfu/m2 for 7 h of exposure). Widespreadcontamination of the patients’ rooms was demon-strated for all three patients. Many of the settleplates were positioned around the periphery ofthe rooms (e.g. window and other ledges, floorand sink), and the majority of these plates werepositive, as were settle plates closer to the patient(e.g. under the bed and on the overbed table).

The rate of MRSA environmental contaminationfrom each patient was significantly reduced by 75e93% using a portable HEPA-filtration unit. Thisreduction was directly related to the rate of airfiltration. The IQAir unit can filter air at a maximumrate of 400 m3/h; however, at this fan speed, theunit is unacceptably noisy for both patients andstaff. The IQAir unit was operated at 235 m3/hfor two patients and at 140 m3/h for one patient.With the volume of the rooms, this equates tobetween 3.6 (Patient B) and 8.5 (Patient A) airchanges/h. None of the rooms had ceiling-mounted supply or extract ventilation, or anen-suite facility with extract ventilation, but thesingle room in the intensive care unit (Patient C)had a wall-mounted extract system that was oncontinuously throughout the study period. Theauthors were unable to determine the rate ofextraction through this extract system, and thusthe additional number of air changes/h withinthe room. However, this extract system alone ap-peared to have little effect on preventing MRSAsurface contamination rates. The mean numberof MRSA cfu/10-h exposure/plate was 23.6 and27.7 on the two days without IQAir filtration, anda highly significant effect was still demonstrablewhen the IQAir filtration was used in addition to

the extract system. There were no other patientson the intensive care unit shown to have MRSAduring the study period for Patient C.

The windows (and doors whenever possible)were kept closed whilst the air-filtration unit wasrunning. However, on a number of occasions onstudy days, the doors were observed to have beenleft open, mainly for patient safety reasons. Thiswould have reduced the effectiveness of the IQAirunits by reducing the number of air recirculations,and it is possible that even greater reductions inMRSA surface contamination rates would havebeen observed in a more tightly controlled clinicalenvironment. No attempt was made to record thenumber of people entering the rooms, or allthe other factors that may have affected eitherthe ventilation or the degree of MRSA dispersal(e.g. bed making), and there is likely to have beensome day-to-day variation in these. However,relatively long settle plate exposure times wereused during the main working day in order to‘capture’ most of these potential variables,and the efficacy of IQAir filtration was demon-strated consistently in different busy clinicalenvironments.

The effectiveness of air filtration may be de-pendent on the positioning of the portable unit.8,9

However, the IQAir units were placed in relativelydifferent positions for each patient (so as not to bein the way of clinical staff) (see Figure 1), and thisdid not appear to alter the overall efficiency of airfiltration. Similarly, Rutala et al. found no differ-ence in the clearance rate of airborne particleswith portable units positioned in different loca-tions within a room.10 Figure 1 clearly shows theeffect of IQAir filtration on MRSA contaminationrates throughout the rooms of Patients A and B.Figure 1(c) could be interpreted as showinga greater effect nearer the machine, but this isnot apparent if the actual numbers of MRSA cfu/10-h exposure/plate (data not shown) are com-pared, and the reduction in the number of MRSAcfu was relatively evenly distributed across theroom.

One possible confounding explanation for theseresults is that the patients shed less MRSA on dayswhen air filtration was used. The rate of dispersalof MRSA may be related to factors such as patientmovement, clothing, medical procedures (e.g.dressing changes, suctioning), bed making, andtopical or systemic MRSA antiseptics or antibi-otics.11 Patient A was ambulatory, but his patternof movement within the room was relatively con-sistent on each study day, and he wore similarclothes. His rate of MRSA shedding decreasedover time, but this was controlled for by

Portable HEPA filtration and MRSA 53

alternating days with and without air filtration,and using adjusted rate ratios in the Poisson re-gression analysis. Both Patients B and C were con-fined to bed, and their clinical conditionsremained largely unchanged throughout the studyperiods. Patient B required regular suctioning viaa tracheostomy, and Patient C required wounddressing changes on alternate days which hap-pened on both study days with air filtration on(see Table I). Whilst study days with and withoutair filtration for Patients B and C were not alter-nated, the rate of MRSA contamination fromPatient B increased as the IQAir fan speed wasdecreased on consecutive days, indicating ongoingdispersal. Topical MRSA treatment was not used inany of the patients in the study period. AlthoughPatient C was treated with intravenous vancomy-cin and gentamicin, his skin wounds remainedMRSA positive. For these reasons, it is consideredto be unlikely that these confounding factors ac-count for the observed results.

Portable HEPA-filtration units have previouslybeen shown to be efficient at removing airborneparticles, and they are sometimes employed inhigh-risk units to reduce risks of invasive aspergil-losis during nearby building demolition or con-struction.12 Published evaluations have generallyused artificially generated aerosols in a test settingto demonstrate that particles of the size of bacte-ria, droplet nuclei or fungal spores are removed ef-ficiently and quickly.8e10 The present study hasdemonstrated that portable HEPA-filtration unitscan significantly reduce the amount of potentialMRSA contamination of horizontal surfaces andequipment within the rooms of patients who areheavy MRSA dispersers.

Although there is no direct proof, there isincreasing evidence that the environment can actas a reservoir for S. aureus, including MRSA, andthat this can pose a risk of cross-infection to pa-tients.2,3,7,13e16 A heavily contaminated environ-ment poses a risk that healthcare workers maycontaminate their hands, gloves and clothing.Boyce et al. found that 65% of nurses had contam-inated their uniform or gowns with MRSA whilstlooking after MRSA patients, and 42% of personnelwho had no direct contact with these patients, buthad touched contaminated surfaces within theirrooms, had contaminated their gloves withMRSA.7 The present study has shown that therate of environmental MRSA contamination canbe reduced significantly by portable air filtration,and this should reduce these cross-infection risks.Studies have also shown that routine and terminalcleaning are not 100% effective in clearing MRSAfrom the environment,3,4 so a measure that

reduces MRSA environmental contamination inthe first place will reduce the consequences of in-effective cleaning. It is envisaged that portableHEPA-filtration units may also be of value in a vari-ety of different clinical settings (e.g. in dressingclinics, for respiratory function tests or on wardswith MRSA patients who cannot be isolated), butfurther work is needed.

In conclusion, placing IQAir portable HEPA-filtration units within MRSA isolation rooms cansignificantly reduce the contamination of environ-mental surfaces with MRSA. Although this cannotreplace standard infection control measures (e.g.isolation, hand hygiene, protective clothing andcleaning), it is likely to reduce cross-infection riskssignificantly and could provide a relatively cost-effective method for enhancing MRSA control.Further evaluation of these units is required withinlarger or more open-plan areas (e.g. intensive careunits), and the relationship between reducingenvironmental contamination and MRSA coloniza-tion/infection rates needs to be established.

Acknowledgements

The authors are grateful to Air Science Ltd forproviding two IQAir H13 Cleanroom units for thisstudy. The authors would also like to thank Dr SarahLewis for helping with the statistical analysis.

References

1. Boyce JM, Pittet D. Guideline for hand hygiene in healthcare settings: recommendations of the Healthcare InfectionControl Practices Advisory Committee and the Hand Hy-giene Task Force. Infect Control Hosp Epidemiol 2003;23(Suppl):1e40.

2. Talon D. The role of the hospital environment in the epide-miology of multi-drug resistant bacteria. J Hosp Infect 1999;43:13e17.

3. Rampling A, Wiseman S, Davis L, et al. Evidence that hospi-tal hygiene is important in the control of MRSA. J HospInfect 2001;49:109e116.

4. French GL, Otter JA, Shannon KP, Adams NMT, Watling D,Parks MJ. Tackling contamination of the hospital environ-ment by MRSA: a comparison between conventionalterminal cleaning and hydrogen peroxide vapour decontam-ination. J Hosp Infect 2004;57:31e37.

5. Ayliffe GAJ. Revised guidelines for the control of methicil-lin-resistant Staphylococcus aureus infection in hospitals.J Hosp Infect 1998;39:253e290.

6. Sheretz RJ, Reagan DR, Hampton KD, et al. A cloud adult:the Staphylococcus aureus-virus interaction revisited. AnnIntern Med 1996;124:539e547.

7. Boyce JM, Potter-Bynoe G, Chenevert C, King T.Environmental contamination due to methicillin-resistantStaphylococcus aureus: possible infection control impli-cations. Infect Control Hosp Epidemiol 1997;18:622e627.

54 T.C. Boswell, P.C. Fox

8. Miller-Leiden S, Lobascio C, Nazaroff WW, Macher JM.Effectiveness of in-room air filtration and dilution ventila-tion for tuberculosis infection control. J Air Waste ManagAssoc 1996;46:869e882.

9. Lawrence JC, Lilly HA, Wilkins MD. Evaluation of a portableair purifier. J Hyg 1981;86:203e208.

10. Rutala WA, Jones SM, Worthington JM, Reist PC,Weber DJ. Efficacy of portable filtration units in reducingaerosolized particles in the size range of Mycobacteriumtuberculosis. Infect Control Hosp Epidemiol 1995;16:391e398.

11. Shiomori T, Miyamoto H, Makishima K, et al. Evaluationof bedmaking-related airborne and surface methicillin-resistant Staphylococcus aureus contamination. J HospInfect 2002;50:30e35.

12. Loo VG, Bertrand C, Dixon C, et al. Control of construction-associated nosocomial aspergillosis in an antiquated

hematology unit. Infect Control Hosp Epidemiol 1996;17:360e364.

13. Moore EP, Williams EW. A maternity hospital outbreak ofmethicillin-resistant Staphylococcus aureus. J Hosp Infect1991;19:5e16.

14. Blythe D, Keenlyside D, Dawson SJ, Galloway A. Environ-mental contamination due to methicillin-resistant Staphy-lococcus aureus. J Hosp Infect 1998;38:67e70.

15. Kumari DN, Haji TC, Keer V, Hawkey PM, Duncanson V,Flower E. Ventilation grilles as a potential source of meth-icillin-resistant Staphylococcus aureus causing an outbreakin an orthopaedic ward at a district general hospital. J HospInfect 1998;39:127e133.

16. Layton MC, Perez M, Heald P, Patterson JE. An outbreak ofmupirocin-resistant Staphylococcus aureus on a dermatol-ogy ward associated with an environmental reservoir. InfectControl Hosp Epidemiol 1993;14:369e375.