The use of adenosine triphosphate bioluminescence to assess the efcacy of a modied cleaning program implemented within an intensive care

setting
Ginny Moore, PhD,a Debbie Smyth, RGN,a Julie Singleton, RGN,b and Peter Wilson, MD, FRCP, FRCPatha London, United Kingdom

Background: A total environmental cleaning system based on microber technology was implemented within 2 intensive care units (ICUs). The efcacy of this modied cleaning program was assessed using adenosine triphosphate (ATP) bioluminescence. Methods: A team of trained hygiene technicians cleaned all near-patient furniture and equipment twice a day using ultramicrober cloths. Every week for 40 weeks, 10 surfaces within a randomly selected bed area were sampled using the 3M Clean-Trace Clinical Hygiene Monitoring System (3M Health Care Ltd, Loughborough, United Kingdom). The ability of the modied cleaning program to reduce surface contamination to acceptable levels was measured against previously proposed benchmark ATP values. Results: In comparison with normal cleaning procedures routinely carried out by the nurses, the modied cleaning program signicantly reduced (P , .001) the ATP readings obtained from surfaces within the near-patient environment. In both ICUs, 95% of surfaces sampled after modied cleaning had relative light unit values of ,500 and were deemed clean. Almost 90% of the surfaces could also be passed using the more stringent benchmark value of 250 relative light units. However, regardless of benchmark value used, the majority of surfaces sampled could also be considered adequately clean prior to them being cleaned by the hygiene technicians. Conclusion: The use of ATP bioluminescence has been proposed as a means to improve the management of hospital cleaning. Use of benchmark values can help continually monitor the efcacy of existing cleaning programs. However, when evaluating novel or new cleaning practices, baseline cleanliness (ie, the level of cleanliness routinely achieved using normal cleaning procedures) must also be taken into consideration, or the efcacy of modied cleaning will be overestimated. Key Words: Adenosine triphosphate bioluminescence; ATP bioluminescence; intensive care unit; cleaning. Copyright 2010 by the Association for Professionals in Infection Control and Epidemiology, Inc. Published by Elsevier Inc. All rights reserved. (Am J Infect Control 2010;38:617-22.)

From the Department of Microbiology, University College London Hospitals NHS Foundation Trust, London, United Kingdoma; and Department of Intensive Care Royal Free Hampstead NHS Trust, London, United Kingdom.b Address correspondence to Ginny Moore, PhD, University College London Hospitals NHS Foundation Trust, Environmental Laboratory, Windeyer Institute of Medical Sciences, 46 Cleveland St, London W1T 4JF, United Kingdom. E-mail: ginny.moore@uclh.nhs.uk. This is an independent report commissioned and funded by the Policy Research Program in the Department of Health. The views expressed are not necessarily those of the Department. A.P.R. Wilson was partly funded by the UCLH/UCL Comprehensive Biomedical Centre, which received a proportion of funding from the Department of Healths NIHR Biomedical Research Centres funding scheme. CuWB50 was provided free of charge by ICICS plc, London. Conicts of interest: None to report. 0196-6553/$36.00 Copyright 2010 by the Association for Professionals in Infection Control and Epidemiology, Inc. Published by Elsevier Inc. All rights reserved. doi:10.1016/j.ajic.2010.02.011

Hospital cleanliness may be linked with the level of hospital-acquired infection. In the United Kingdom, a number of initiatives have been launched that aim to improve the standards of cleaning, yet cleaning programs continue to be criticized for their lack of efcacy and poor management.1 Good cleaning management should incorporate some means of monitoring cleaning efcacy,2 and, to comply with the National Specications for Cleanliness,3 National Health Service trusts must now not only demonstrate that their hospitals are clean but how and to what standard they are kept clean. Cleaning audits are a commonly employed means of assessing the quality of hospital cleaning programs. However, many such inspections use visual assessment as a performance criterion. Whereas the visual inspection of surfaces can reveal gross deciencies caused by the presence of visible blood, body spillages, dust, and food, it has been well documented that visual assessment is a poor indicator of cleaning efcacy.1,4-6 Realistically, hospitals require information on surface cleanliness that extends far beyond the sensitivity and subjectivity of a visual test. 617

618

Moore et al.

American Journal of Infection Control October 2010

Adenosine triphosphate (ATP) is present in all types of organic material (eg, food, bacteria, bodily uids), and its detection, via ATP bioluminescence, provides an indication of total (ie, microbial and nonmicrobial) surface contamination. The ability to obtain results in real time has meant that the technique has been used to great effect within the food industry where its regular and frequent use can provide management with data regarding trends in level of hygiene. Rapid results also mean that any sudden problems associated with a cleaning regimen can be identied and rectied rapidly. The use of ATP bioluminescence within the health care environment has been more limited. However, in the United Kingdom, the Clean-Trace Clinical Hygiene Monitoring System (3M Health Care Ltd, Loughborough, United Kingdom) has recently been awarded Rapid Review Panel Recommendation 1.7 The Rapid Review Panel was set up by the Department of Health to assess new and novel products which may help infection prevention and control. Technologies that receive a Rapid Review Panel 1 recommendation have shown, via basic research and development, validation, and in-use evaluations, benets that should be available to National Health Service bodies. The panel concluded that the Clean-Trace ATP system (3M Health Care Ltd) can be used to measure cleanliness within a clinical setting.7 However, such single-shot ATP systems are unable to distinguish between microbial and nonmicrobial ATP, and, as such, they should not be considered microbiologic test methods. Several studies have demonstrated poor correlation between results obtained using ATP bioluminescence and those using traditional microbiologic surface sampling techniques.2,6 Such results are not unexpected and demonstrate how the relative proportion of microbial and nonmicrobial contamination can vary depending on the type of surface contamination present. Nonetheless, whereas ATP bioluminescence values cannot be directly equated to the level of surface-associated bacteria unacceptable levels of ATP have been shown to correlate with unacceptable levels of microorganisms.5 Thus, the ability of a cleaning program to reduce total organic soil (ie, ATP levels) to an acceptable level can be used as a marker for its ability to reduce microbial contamination. Any cleaning program unable to deliver appropriately low ATP levels is unlikely to be able to remove potential pathogens from a surface and, as such, is unlikely to be t for purpose.8 When implementing ATP bioluminescence, therefore, it is essential to initially establish acceptable ATP levelsATP levels that are indicative of effective cleaning. Benchmark ATP valuesvalues that can be consistently achieved following best practice cleaninghave been proposed. Grifth et al4 sampled a variety of surfaces located within a number of different ward

environments and concluded that, when using the Clean-Trace ATP system (3M Health Care Ltd), 500 relative light units (RLU) is a realistic benchmark value. However, the authors acknowledged that a very basic cleaning schedule was in existence and hypothesized that it would be possible, through the implementation of a well constructed, validated cleaning program, to consistently achieve lower bioluminescence readings. Indeed, on the basis of results obtained during a more recent study, which also utilized the CleanTrace ATP system, Lewis et al8 proposed a more stringent benchmark value of 250 RLU. Benchmarking, by providing objective and attainable cleaning standards, enables the performance of existing practices to be continually assessed helping to improve both the efcacy and quality of cleaning programs. It has also been suggested that benchmark values can be used to evaluate novel or modied cleaning practices.8 The following study was conducted within the critical care setting. In the United Kingdom, near-patient items of clinical equipment are not cleaned by domestic staff but by nurses who are unsupervised and often do not see cleaning as part of their responsibility. Consequently, cleaning may be infrequent and/or ineffective. To enhance normal cleaning procedures, a modied cleaning regimen was implemented within 2 intensive care units and its efcacy assessed by means of ATP bioluminescence. The ability of the modied cleaning program to reduce surface contamination to acceptable levels was measured against previously proposed benchmark ATP values.

METHODS Setting and normal cleaning procedures

This study was conducted within the general medical-surgical intensive care units (ICU) of 2 central London teaching hospitals. At both hospitals, the nurses were expected to clean the bed area and associated equipment once a shift. The normal damp dusting procedure implemented within each ICU differed; one unit utilized Actichlor Plus (a detergent with chlorine releasing agent [1000 ppm]; Ecolab, Swindon, United Kingdom), the other a 70% alcohol spray.

Modied cleaning regimen

A team of hygiene technicians was specically employed and trained to clean all near-patient furniture (eg, bed rails, storage trolleys, chart tables) and equipment (eg, syringe drivers, ventilators, monitors) with ultramicrober (UMF) cloths (Johnson-Diversey UK Ltd, Northampton, United Kingdom). The increased mechanical energy generated by microber strands can improve cleaning efcacy without the use of

www.ajicjournal.org Vol. 38 No. 8

Moore et al.

619

detergent.9,10 However, there are concerns that microber cloths can act as vehicles for the spread of nosocomial pathogens.11,12 During this study, to prevent the redistribution of live organisms during the cleaning process, the cloths were presoaked in a copper-based biocidal formulation (CuWB50; ICICS plc, London, United Kingdom).13 Each bed area was divided into 4 zones with a minimum of 1 cloth being used to clean all furniture and equipment located within each zone. After use, all cloths were laundered in a 928C wash cycle, tumble dried, and reused within 24 hours.

Assessment of cleaning efcacy

The ability of this modied cleaning regimen to effectively clean the near-patient environment was assessed by means of ATP bioluminescence. Every week for 40 weeks, 10 surfaces within a randomly selected bed area were sampled using Clean-Trace Surface ATP swabs (3M Health Care Ltd) both before and within 10 minutes of the modied clean being completed. Although variation in surface size and shape prevented the standardization of surface area sampled, the same area of each test surface was sampled on both occasions. Each swab was activated in accordance with the manufacturers instructions and placed in a Clean-Trace NG Luminometer (3M Health Care Ltd), and the reading in RLU was recorded. The hygiene technicians were blinded to both the bed area and surfaces sampled, and the results obtained were compared with previously proposed benchmark values.

Surface ATP swab. Immediately prior to swab activation, 10 mL of the ATP solution was pipetted onto the middle of the swab bud. The device was activated in accordance with the manufacturers instructions, and readings were taken using the Clean-Trace NG Luminometer. This assay was performed using 10 replicate test areas and then repeated substituting the sterile distilled water with 0.1 mL of the copper-based biocidal formulation. The effect of CuWB50 on the bioluminescence light signal was calculated using the following formula: ((mean RLU value obtained from control [water] assay 2 mean RLU value obtained from test [CuWB50] assay)/mean RLU value obtained from control assay) 3 100.

RESULTS
Before being cleaned by the hygiene technicians, the mean level of contamination present on surfaces routinely cleaned with Actichlor Plus (Ecolab) and 70% alcohol equated to 584 RLU and 495 RLU, respectively. In both cases, the modied cleaning regimen signicantly reduced (P , .001) the level of surface contamination. Overall, the mean ATP reading obtained from surfaces within the near-patient environment (n 5 396) fell signicantly (P , .001) from 534 RLU before cleaning to 122 RLU following modied cleaning. Almost 95% of surfaces cleaned by the hygiene technicians had RLU values of less than 500 (Fig 1) and, thus, in accordance to the ATP benchmark suggested by Grifth et al4 were passed and considered clean. Almost 90% of the surfaces could also be passed using the more stringent pass/fail benchmark value of 250 RLU8 (Fig 1). Nonetheless, the level of surface contamination detected on 7 surfaces (2%) equated to more than 1000 RLU. In comparison, 79% and 58% of surfaces sampled prior to modied cleaning had RLU readings of ,500 and ,250, respectively; 10% of surfaces had RLU readings of .1000 (Fig 2). The results of the laboratory-based trial suggested that sampling a surface in the presence of in-use concentrations of CuWB50 would quench the ATP light signal by approximately 9.5% (Table 1).

Effect of biocide on ATP bioluminescent light signal

Several studies have demonstrated that a number of commonly used cleaning chemicals, when applied at in-use concentrations, can cause the quenching or, in some cases, the enhancement of the ATP light signal.14,15 During the current study, a laboratory-based trial was conducted to ensure that the biocide incorporated within the modied cleaning regimen did not adversely affect the ATP bioluminescence assay and contribute to erroneously high or low RLU readings. A stainless steel surface marked with 10 cm 3 10-cm squares was sanitized using a previously validated cleaning protocol.16 An in-use concentration of the copper-based biocide (150 mg/L of elemental copper) was prepared by adding 50 mL of stock CuWB50 solution (30,000 mg/L of elemental copper; ICICS plc) to 10 mL of distilled water. A 1027 mol/L solution of ATP was prepared by reconstituting a tablet of freezedried ATP (Clean-Trace Surface Positive Control ATP; 3M Health Care Ltd) with 500 mL of sterile distilled water. 0.1 mL of sterile distilled water was spread over a 100-cm2 test area and sampled using a Clean-Trace

DISCUSSION
ATP bioluminescence is used to great effect within the food industry to continually monitor surface cleanliness. Its use within the health care setting is less well established, but the technique is gaining acceptance, particularly when used to assess the efcacy of existing or modied cleaning programs.1,2,4-6,8,17,18 Modied cleaning programs are likely to include new or novel cleaning solutions, which, because of their similarity to chemicals incorporated within the ATP assay, may either degrade or have an additive effect on the ATP

620

Moore et al.

American Journal of Infection Control October 2010

94.69% 89.89% 90 80 100

225 200 175

Cumulative Percentage Frequency

70 150 Number of sites 60 125 50 100 40 75 50 25 0

0 50 100 150 200 250 300 350 400 450 500 550 600 650 700 750 800 850 900 950

30 20 10 0
1000 >1000

ATP reading (Relative Light Units)

Fig 1. ATP levels obtained from surfaces within the near-patient environment after they had been cleaned using a modied cleaning regimen. The percentage of surfaces passed as clean using 2 different benchmark values (250 RLU; 500 RLU- - - - -) is also illustrated.
100 90 78.74% 80 70 80 70
57.76%

100 90

Cumulative Percentage Frequency

Number of sites

60 50 40 30 20 10 0
0 50 100 150 200

60 50 40 30 20 10 0

250

300

350

400

450

500

550

600

650

700

750

800

850

900

950

1000 >1000

ATP reading (Relative Light Units)

Fig 2. ATP levels obtained from surfaces within the near-patient environment before modied cleaning. The percentage of surfaces passed as clean using 2 different benchmark values (250 RLU; 500 RLU- - - - -) is also illustrated. measurement.14,15 The modied cleaning program assessed during this investigation incorporated UMF cloths premoistened with a copper-based biocidal formulation. Although this novel biocide was shown to quench the ATP light signal by approximately 9.5%, this degree of quenching was not statistically signicant (P . .1), and any reduction in ATP values brought about by modied cleaning could be considered genuine. The efcacy of the modied cleaning program was assessed over a 40-week period. Once its initial effect on ATP levels had been determined, continual assessment provided assurance that the laundering process was not having a detrimental effect on the UMF cloths and that the hygiene technicians were consistent in their cleaning technique and intensity. The results imply that the modied cleaning program was and continued to be effective in removing surface bioburden. Nonetheless, whereas modied cleaning signicantly reduced ATP levels, approximately 79% and 58% of surfaces sampled before modied cleaning had RLU readings of ,500 and ,250, respectively, (Fig 2) and therefore could be considered adequately clean prior to being cleaned by the hygiene technicians. Thus, whereas previously proposed benchmark values can provide useful standards when assessing the efcacy of existing cleaning programs, when evaluating modied cleaning practices, the baseline

www.ajicjournal.org Vol. 38 No. 8

Moore et al.

621

Table 1. The effect of the copper-based biocidal formulation CuWB50 on the ATP bioluminescence light signal
Relative light units Control assay: surface swabbed in presence of distilled water 4507 3142 5101 5328 4900 5147 1856 2440 3200 3537 Mean 5 3916 Test assay: surface swabbed in presence of CuWB50 4604 2786 1994 2840 2722 2536 6935 2569 2731 5696 Mean 5 3541

NOTE. In-use concentration of CuWB50 is 150 mg/L of elemental copper.

cleanliness of the environment (ie, the level of cleanliness routinely achieved using normal cleaning procedures) must also be taken into consideration. During this study, if the ATP levels achieved following modied cleaning had been compared directly with either of the previously proposed benchmark values, the efcacy of the modied cleaning protocol would have been overestimated. Although 90% to 95% of surfaces sampled were deemed clean (Fig 1), 20% of those surfaces considered unacceptable prior to being cleaned by the hygiene technicians were still unacceptable after modied cleaning. Those surfaces least likely to be cleaned to an acceptable standard included bed rails and storage trolley drawer handles. This study supports the ndings of Lewis et al8 and suggests that, within an ICU, 250 RLU is an achievable and appropriate ATP benchmark. However, this target value may not be appropriate for all hospital environments. ATP measurements taken from surfaces within non-ICU wards are generally higher than those taken from surfaces within the ICU,7 and it has been suggested that, in full and busy medical and/or surgical wards, ,500 RLU may be a more achievable cleaning standard.6 There is a danger that, if benchmark values are set too low, hospitals may nd themselves cleaning unnecessarily, wasting time, money, and energy. Rather than treating all wards as a single entity, different ward types may need to be treated as separate environments and cleaning targets adjusted accordingly. Benchmark values may also need to be revised if alternative ATP systems are used. The Clean-Trace ATP system (3M Health Care Ltd) differs signicantly from the Hygiena system (Hygiena International Ltd, Watford, UK) in terms of reagent chemistry, light detection system, and the scale units used for presenting the

results17; and, in studies utilizing the latter system, ATP levels as low as 100 RLU have been interpreted as being unacceptably high.5 Results obtained using different ATP systems cannot be considered interchangeable, highlighting the importance of a documented cleaning assessment protocol preventing in-coming managerial staff from implementing their preferred ATP system. Previously developed baseline data become worthless if new testing protocols do not provide equivalent results.19 The results of the current study suggest that effective cleaning of an ICU is likely to consistently reduce ATP levels to very low levels. The modied cleaning program was more effective in removing surface bioburden than the normal cleaning procedures routinely carried out by the nurses. However, it is acknowledged that it is not possible to state whether this improvement was due to the extra attention and diligence given to cleaning by the trained hygiene technicians and/or the increased efcacy of any of the products used. It must also be appreciated that the sensitivity of the ATP assay is such that, in the absence of detectable nonmicrobial ATP, the presence of low levels of microorganisms may be overlooked.16 Meticillin-resistant Staphylococcus aureus, glycopeptide-resistant enterococci, and Enterobacteriaceae have all been isolated from surfaces passed as clean using ATP bioluminescence.5 Thus, rather than being assumed from the results of nonmicrobiologic sampling methods (eg, ATP bioluminescence), if the microbiologic status of a surface is to be assured, assessment must also involve some form of microbiologic testing. During the current study, dip slides (ie, contact plates) were also used to sample the near-patient environment, and the effect of the modied cleaning program on local contamination levels has been described.20 Visual assessment, ATP bioluminescence, and microbiologic sampling methods clearly assess different parameters, and, thus, no single assay procedure can completely characterize the contamination on a surface. Instead, the different methods should be seen to complement each other, and the implementation of an integrated cleaning assessment strategy4,6,8 should enable hospitals to ensure that their cleaning and disinfection procedures are and continue to be effective.
The authors thank Francois Senga, Louise Walker, and Erica Balla.

References
1. Cooper RA, Grifth CJ, Malik RE, Obee P, Looker N. Monitoring the effectiveness of cleaning in four British hospitals. Am J Infect Control 2007;35:338-41. 2. Grifth CJ, Obee P, Cooper RA, Burton NF, Lewis M. The effectiveness of existing and modied cleaning regimens in a Welsh hospital. J Hosp Infect 2007;65:352-9.

622

Moore et al.

American Journal of Infection Control October 2010 12. Bergen LK, Meyer M, Hg M. Spread of bacteria on surfaces when cleaning with microbre cloths. J Hosp Infect 2009;71:132-7. 13. Gant VA, Wren MWD, Rollins MSM, Jeanes A, Hickok SS, Hall TJ. Three novel highly charged copper-based biocides: safety and efcacy against healthcare-associated organisms. J Antimicrob Chemother 2007;60:294-9. 14. Simpson WJ, Hammond JRM. The effect of detergents on rey luciferase reactions. J Biolum Chemilum 1991;6:97-106. 15. Velazquez M, Feirtag JM. Quenching and enhancement effects of ATP extractants, cleansers and sanitizers on the detection of the ATP bioluminescence signal. J Food Prot 1997;60:799-803. 16. Moore G, Grifth C, Fielding L. A comparison of traditional and recently developed methods for monitoring surface hygiene within the food industry: a laboratory study. Dairy Food Environ Sanit 2001;21:478-88. 17. Anderson BM, Rasch M, Kvist J, Tollefsen T, Lukkassen R, Sandvik L, et al. Floor cleaning: effect on bacteria and organic materials in hospital rooms. J Hosp Infect 2009;71:57-65. 18. Boyce JM, Havill NL, Dumigan DG, Golebiewski M, Balogun O, Rizvani R. Monitoring the effectiveness of hospital cleaning practices by use of an adenosine triphosphate bioluminescence assay. Infect Control Hosp Epidemiol 2009;30:678-84. 19. Swanson KMJ, Anderson JE. Industry perspectives on the use of microbial data for hazard analysis and critical control point validation and verication. J Food Prot 2000;63:815-8. 20. Moore G, Smyth D, Singleton J, Jackson R, Bellingan G, Wilson P, et al. Implementation of an enhanced cleaning regimen within two intensive care units: the effect on local contamination rates of MRSA and other pathogens. 19th Annual Scientic Meeting of The Society for Healthcare Epidemiology of America. San Diego, March 19-22, 2009.

3. The National Patient Safety Agency. The national specications for cleanliness in the NHS: a framework for setting and measuring performance outcomes. April 2007. The National Patient Safety Agency. London, UK. 4. Grifth CJ, Cooper RA, Gilmore J, Davies C, Lewis M. An evaluation of hospital cleaning regimes and standards. J Hosp Infect 2000;45: 19-28. 5. Willis C, Morley R, Westbury J, Greenwood M, Pallett A. Evaluation of ATP bioluminescence swabbing as a monitoring and training tool for effective hospital cleaning. Br J Infect Control 2007;8:17-21. 6. Sherlock O, OConnell NO, Creamer E, Humphreys H. Is it really clean? An evaluation of the efcacy of four methods for determining hospital cleanliness. J Hosp Infect 2009;72:140-6. 7. Department of Health/NHS Purchasing and Supply Agency. The healthcare associated infections (HCAI) technology innovation programme: showcase hospitals reports No. 2: the 3M Clean-Trace Clinical Hygiene Monitoring System. July 27, 2009. Available from: http://www.clean-safecare.nhs.uk/index.php?pid584. Accessed November 28, 2009. 8. Lewis T, Grifth C, Gallo M, Weinbren M. A modied ATP benchmark for evaluating the cleaning of some hospital environmental surfaces. J Hosp Infect 2008;69:156-63. rgensen O, Schneider T. Micro-bre and ultra-micro9. Nilsen SK, Dahl I, Jo bre cloths, their physical characteristics, cleaning effect, abrasion on surfaces, friction and wear resistance. Build Environ 2002;37:1373-8. berg A10. Pesonen-Leinonen E, Redsven I, Kuisma R, Hautala M, Sjo M. Cleaning efcacies of mop cloths on oor coverings. Tenside Surf Det 2003;40:80-6. 11. Moore G, Grifth C. A laboratory evaluation of the decontamination properties of microbre cloths. J Hosp Infect 2006;64: 379-85.