Evaluation of two methods for monitoring surface cleanliness—ATP bioluminescence and traditional hygiene swabbing (original) (raw)

Evaluation of two methods for monitoring surface cleanliness-ATP bioluminescence and traditional hygiene swabbing

C. A. Davidson,† C. J. Griffith,* A. C. Peters and L. M. Fielding
Food Safety Research Group, University of Wales Institute Cardiff, Colchester Avenue Campus, Colchester Avenue, Cardiff CF3 7XR, UK

Received 4 February 1998; revised 28 May 1998; accepted 29 June 1998

Abstract

The minimum bacterial detection limits and operator reproducibility of the Biotrace Clean-Trace ® { }^{\text {® }} Rapid Cleanliness Test and traditional hygiene swabbing were determined. Areas (100 cm2)\left(100 \mathrm{~cm}^{2}\right) of food grade stainless steel were separately inoculated with known levels of Staphylococcus aureus (NCTC 6571) and Escherichia coli (ATCC 25922). Surfaces were sampled either immediately after inoculation while still wet, or after 60 min when completely dry. For both organisms the minimum detection limit of the ATP Clean-Trace ® { }^{\text {® }} Rapid Cleanliness Test was 104cfu/100 cm2(p<0.05)10^{4} \mathrm{cfu} / 100 \mathrm{~cm}^{2}(p<0.05) and was the same for wet and dry surfaces. Both organism type and surface status (i.e. wet or dry) influenced the minimum detection limits of hygiene swabbing, which ranged from 105cfu/100 cm210^{5} \mathrm{cfu} / 100 \mathrm{~cm}^{2} to >107cfu/100 cm2>10^{7} \mathrm{cfu} / 100 \mathrm{~cm}^{2}. Hygiene swabbing percentage recovery rates for both organisms were less than 0.1%0.1 \% for dried surfaces but ranged from 0.33%0.33 \% to 8.8%8.8 \% for wet surfaces. When assessed by six technically qualified operators, the Biotrace CleanTrace ®{ }^{\circledR} Rapid Cleanliness Test gave superior reproducibility for both clean and inoculated surfaces, giving mean coefficients of variation of 24%24 \% and 32%32 \%, respectively. Hygiene swabbing of inoculated surfaces gave a mean CV of 130%130 \%. The results are discussed in the context of hygiene monitoring within the food industry. Copyright © 1999 John Wiley & Sons, Ltd.

KEYWORDS: ATP bioluminescence; hygiene swabbing; HACCP; monitoring; surface cleanliness

INTRODUCTION

The Hazard Analysis Critical Control Point (HACCP)based approach to food safety management is now widely accepted (1). HACCP-based systems should conform to the internationally agreed principles and guidelines outlined by the Codex Alimentarius Commission (2). For a HACCP-based approach to be effective, it will depend upon the accurate identification of hazards, and the suitability of the control measures and monitoring procedures used. One such hazard might be the contamination of food contact surfaces, with cleaning being designated as a control measure.

As yet, no standard protocol has been adopted by industry for surface hygiene monitoring (3). A recent (1996) unpublished survey of 500 food manufacturing businesses in the UK, carried out by these authors, has corroborated these findings. The survey revealed that some businesses sampled wet surfaces whilst others sampled surfaces that were dry. Of the methods currently available for monitoring surface cleanliness, it was found

[1]that 48%48 \% of respondents used swabbing followed by bacterial culture, while 27%27 \% used ATP bioluminescence.

Traditionally, the assessment of food contact surface cleanliness involved using microbiological methods, including hygiene swabbing and agar contact methods (4). These methods, which are simple to use, can provide either qualitative or quantitative information on the microbial load present. Microbiological methods do, however, require incubation periods of 24−48 h24-48 \mathrm{~h}, which makes cleanliness testing retrospective (5), and therefore inappropriate for use as a rapid monitoring tool required for HACCP (1).

Over the last decade, ATP bioluminescence has become increasingly adopted for monitoring surface cleanliness, and its use is predicted to increase substantially in the near future (1). This method provides a realtime estimate of total surface cleanliness, including the presence of organic debris and microbial contamination. The ability to provide results within minutes, as opposed to days for microbiological testing, enables ATP bioluminescence to be used as a monitoring method within HACCP. While traditionalists might argue that the primary concern should be the detection of microbial contamination only, the presence of organic debris, besides indicating inadequate cleaning, provides a source of nutrients which will support the growth of even very low numbers of microorganisms (5).

1. *Correspondence to: C. J. Griffith, Food Safety Research Group, University of Wales Institute Cardiff, Colchester Avenue Campus, Colchester Avenue, Cardiff CF3 7XR, UK.
   †\dagger Present address: School of Environmental Sciences and Land Management, University College Worcester, Henwick Grove, Worcester WR2 6AJ, UK. Email: c.davidson@worc.ac.uk ↩︎

Several studies have compared the results gained from microbiological and ATP Bioluminescence based methods for assessing in situ surface cleanliness. Some studies have illustrated a good correlation between the results for ATP and microbiological methods (5-7), others have illustrated a poor correlation (1,8)(1,8). This disparity in findings is not surprising given the fact that in situ, details of the nature of surface contamination are unknown. High ATP readings, for example, might be the result of varying combinations of ATP derived from food and microbial origin. It may be important, however, that people monitoring surface cleanliness using bioluminescence are able to detect low numbers of microorganisms in the absence of ATP derived from food residues.

The reproducibility of any hygiene test method is defined as the ability of the swab, reagents and equipment, when used by any operative under the same sampling conditions, to produce the same end result given a consistent level of initial bioburden. Reproducibility is, therefore, a vital factor in the monitoring of control measures and the setting of target values and critical limits, especially given the increasing use of results for trend analysis (1).

The aim of the work reported in this paper was to determine the minimum bacterial detection limits and reproducibility of hygiene swabbing and ATP bioluminescence when used to sample stainless steel surfaces contaminated with known levels of microbial bioburden.

MATERIALS AND METHODS

Cultures

Cultures of Escherichia coli (ATCC 25922) and Staphylococcus aureus (NCTC 6571) were grown in Nutrient Broth No. 2 (Oxoid, Basingstoke, UK) in unshaken batch culture volumes of 10 mL under aerobic conditions at 37∘C37^{\circ} \mathrm{C} for 18 h . Serial dilutions of each culture, prepared in Maximum Recovery Diluent (MRD; Oxoid), were used to inoculate a food-grade stainless steel surface.

ATP bioluminescence test method

A range of ATP bioluminescence equipment and reagents is available, but previous work (1) had shown that the Biotrace Clean-Trace ®{ }^{\circledR} Rapid Cleanliness Test was typical of those available in terms of its minimum detection limit and reproducibility, and was found to be the easiest to use.

The test consists of a swab device containing a dacron bud, pre-moistened with a cationic agent to release ATP from any intact food or microbial cells, and to aid soil removal from surfaces. Firefly reagent released on
activation of the device reacts with ATP collected on the swab bud and produces light. The intensity of light is proportional to the amount of ATP and therefore the degree of contamination.

Hygiene swabs and swabbing protocol

Sterile cotton hygiene swabs with wooden applicator sticks (Technical Service Consultants Ltd, Heywood, UK), pre-moistened in MRD immediately before use, were used for microbiological sampling.

In all experiments, including those evaluating operator reproducibility, a standard surface swabbing protocol was used. The protocol involved ensuring that the swab bud came into contact with the entire 100 cm2100 \mathrm{~cm}^{2} surface area, that the swab was rotated constantly during swabbing, and that each surface was swabbed in two directions at 90∘90^{\circ} to each other. Swabs were held by their handles, and not by their applicator sticks, in an attempt to standardize the amount of pressure applied to the swab during sampling.

Surface preparation

Prior to inoculation, the stainless steel surface, marked with 100 cm2100 \mathrm{~cm}^{2} areas, was sanitized for 30 min using a 1:80 dilution of Bioscan containing <5%<5 \% cationic surfactants, 5−15%5-15 \% non-ionic surfactants and <5%<5 \% phosphoric acid (Henkel Ecolab Ltd, Swindon, UK), which acts through disrupting cell membrane integrity. The surface was then cleaned using an ‘in house’ validated cleaning protocol. This involved washing with hot water and detergent ( <5%<5 \% amphoteric, 5−15%5-15 \% non-ionic, and 15−30%15-30 \% anionic surfactants), applying kinetic energy for 2 min using a new rayon cloth, thoroughly rinsing with hot water to remove all traces of detergent, and a final rinse with boiling water before being left to air dry at room temperature. This was demonstrated to give negative microbiological results and ATP bioluminescence readings consistently below 100 RLU using the Biotrace Unilite Xcel and Clean-Trace ®{ }^{\circledR} Rapid Cleanliness Test.

Surface inoculation, sampling and organism cultivation

0.1 mL of each culture (107cfu)\left(10^{7} \mathrm{cfu}\right) and dilutions giving inoculation levels of 106−102cfu10^{6}-10^{2} \mathrm{cfu} were inoculated onto each of 40100 cm240100 \mathrm{~cm}^{2} areas of stainless steel and spread over the entire surface area to within 2 mm of the edges using a sterile plastic spreader. Of the 40 inoculated surfaces, 20 were sampled while wet, and 20 when dry. Ten surfaces were sampled using the Biotrace Clean-

Table 1. Minimum bacterial detection limits of two hygiene monitoring methods on wet and dry inoculated stainless steel surfaces

Detection limits given in brackets indicate consistent recovery of less than 30 cfu/plate.

Trace ®{ }^{\circledR} Rapid Cleanliness Test and 10 with hygiene swabbing, for the wet and dry inoculated surfaces.

Dry surfaces were achieved through allowing the inoculum applied to the surface to remain resident on the stainless steel for 60 min at an ambient temperature of approximately 22∘C22^{\circ} \mathrm{C} prior to sampling. This resulted in no visible liquid culture remaining on the surface at the time of swabbing.

Organisms recovered from inoculated surfaces using hygiene swabs were either released into 10 mL tubes of MRD by rotamixing ( 5 s ) and duplicate 0.1 mL nutrient agar (NA) spread plates prepared, or swabs were streaked directly over the surface of NA plates in two directions in order to maximize release of the recovered organisms from the swabs. All NA plates were incubated at 37∘C37^{\circ} \mathrm{C} for 24 h .

Determination of minimum detection limits

For ATP bioluminescence test results, data from inoculated surfaces were compared with control data from a cleaned surface using a 1-tailed Student’s tt test to estimate the significance of any difference. The minimum detection limit was recorded as the inoculum level, which resulted in a significant (p<0.05)(p<0.05) difference.

For plate count data from hygiene swabs the minimum detection limit was noted as that level of inoculum which resulted in consistent counts of between 30-300 colonies/ plate (ISO standard). In addition, minimum bacterial detection limits are indicated in brackets where counts of less than 30 cfu/plate were recorded consistently for all 10 replicates. Percentage recovery rates were calculated for both organisms from wet and dry inoculated surfaces.

Determination of extracellular ATP levels

Extracellular ATP levels were determined to ensure that the ATP signal being detected was not significantly affected by ATP present in the growth medium. Eighteen hour cultures of each organism were divided into two 5 mL portions and centrifuged at 2000 rpm for 15 min . In
order to ensure that both portions of each culture received identical processing, the supernatant from each was removed. The supernatant from one portion was replaced and the microbial cells resuspended. The supernatant from the second portion was filtered through a 0.45μ m0.45 \mu \mathrm{~m} filter (Whatman, Maidstone, UK) to remove residual cells and retained for ATP analysis. 5 mL of quarter-stength Ringer’s solution was then used to resuspend the microbial cells of the second portion of each bacterial culture.

For the culture suspended in nutrient broth, the culture suspended in quarter strength Ringer’s solution, and the filtered supernatant for each organism, 10 Clean Trace ®{ }^{\circledR} Rapid Cleanliness Test swabs were inoculated with 0.1 mL of a 10−210^{-2} dilution of the suspended microbial cells, or the filtered supernatant. All swabs were assayed as described previously. The mean ATP reading for the filtered supernatant was calculated as a percentage of the mean ATP reading for the culture suspended in Ringer’s solution.

Determination of operator reproducibility

Studies designed to assess operator reproducibility involved six technically qualified operatives swabbing each of 10100 cm210100 \mathrm{~cm}^{2} surfaces which were either clean or marginally unclean through inoculation with low but detectable numbers (105cfu/100 cm2)\left(10^{5} \mathrm{cfu} / 100 \mathrm{~cm}^{2}\right) of E. coli. All operatives received training in both the use of ATP bioluminescence and hygiene swabbing prior to surface sampling. Experiments were repeated with operatives using both the Biotrace Clean-Trace ®{ }^{\circledR} Rapid Cleanliness Tests and conventional hygiene swabs. The latter were streaked directly onto NA plates and incubated for 24 h at 37∘C37^{\circ} \mathrm{C}.

RESULTS

Table 1 illustrates the minimum bacterial detection limits achieved by both test methods when used to sample wet and dry surfaces inoculated with either SS. aureus or EE.

Table 2. Percentage recovery rates for S. aureus and E. coli using hygiene swabbing

coli. The minimum bacterial detection limit for the Clean-Trace 20{ }^{20} Rapid Cleanliness Test was found to be 104cfu/100 cm210^{4} \mathrm{cfu} / 100 \mathrm{~cm}^{2} when used to sample either wet or dry surfaces, and was found to be the same for both test organisms. Extracellular ATP levels in a 10−210^{-2} dilution of the culture growth medium, that was used for surface inoculations, were found to account for between 0.47%0.47 \% and 1.7%1.7 \% of the ATP signal detected. For hygiene swabbing, it was found that the minimum detection limit was influenced by a number of factors, including whether surfaces were wet or dry, the target organism, and the method of cultivation used. The lowest minimum detection limit for SS. aureus using hygiene swabbing was 102cfu/100 cm210^{2} \mathrm{cfu} / 100 \mathrm{~cm}^{2} which was gained from sampling wet inoculated surfaces followed by direct streaking of the swab onto nutrient agar. This minimum detection limit increased to 105cfu/100 cm210^{5} \mathrm{cfu} / 100 \mathrm{~cm}^{2} when detecting the same organism when fully dried on the surface. The minimum detection limits for wet and dry inoculated surfaces were 10310^{3} and 106cfu/100 cm210^{6} \mathrm{cfu} / 100 \mathrm{~cm}^{2}, respectively, when 0.1 mL spread plates were prepared from recovery diluents. The lowest minimum detection limit for E. coli was 102cfu/100 cm210^{2} \mathrm{cfu} / 100 \mathrm{~cm}^{2} gained from sampling wet inoculated surfaces followed by direct streaking onto nutrient agar. This minimum detection limit increased to 107cfu/10^{7} \mathrm{cfu} / 100 cm2100 \mathrm{~cm}^{2} when the organism had been allowed to dry on the surface before sampling. Using 0.1 mL spread plates for cultivation, the minimum detection limits for E. coli on wet and dry inoculated surfaces were 10310^{3} and >107>10^{7} cfu/100 cm2\mathrm{cfu} / 100 \mathrm{~cm}^{2}, respectively. The minimum detection limits for each organism using the selected methods were confirmed through repeating each experiment using the lowest level of inoculum at which detection was achieved.

Hygiene swabbing percentage recovery rates for both
organisms are given in Table 2. Recovery rates were found to be higher for surfaces sampled when wet and, on all but one occasion, when the method of cultivation used was direct streaking onto nutrient agar. For S. aureus, percentage recoveries were 0.33%0.33 \% and 8.8%8.8 \% when sampled from wet inoculated surfaces using 0.1 mL spread plates and direct streaking, respectively. From dry inoculated surfaces, these recovery rates decreased to 0.05%0.05 \% and 0.001%0.001 \% for 0.1 mL spread plates and direct streaking. Percentage recovery rates for E. coli, were lower than those for S. aureus, with a percentage recovery of 7.7%7.7 \% being gained for sampling wet surfaces followed by direct streaking. On dry inoculated surfaces using direct streaking for cultivation, the recovery rate was 0.009%0.009 \%. For sampling wet and dry inoculated surfaces, followed by preparation of 0.1 mL spread plates from recovery diluents, recovery rates were found to be 0.25%0.25 \% and 0%0 \%, respectively.

Table 3 provides mean coefficients of variation (CV) and 95%95 \% confidence limits for ATP bioluminescence and hygiene swabbing when used to sample clean and marginally unclean stainless steel surfaces. A mean CV of 24%24 \% was gained for sampling clean sanitized stainless steel using the Biotrace Clean-Trace 20{ }^{20} Rapid Cleanliness Test. No data could be obtained from hygiene swabbing of cleaned surfaces as no colony forming units were recovered. For marginally unclean surfaces, the ATP test method gave a mean CV of 32%32 \%. When sampling marginally unclean surfaces using hygiene swabbing, the mean CV was found to be 130%130 \%.

DISCUSSION

The purpose of the work reported was to evaluate two
Table 3. Mean coefficients of variation (CV) and 95% confidence limits gained from six operatives when using ATP bioluminescence and hygiene swabbing to sample clean and marginally unclean stainless steel surfaces

Microbiological testing of clean surfaces gave no counts.

methods widely used by the food industry for assessing food contact surface cleanliness. The minimum bacterial detection limits of both the Biotrace Clean-Trace 20{ }^{20} Rapid Cleanliness Test and traditional hygiene swabbing, when used to sample wet and dry surfaces inoculated with pure cultures, were determined. The minimum detection limit of the ATP bioluminescence assay was found to be the same for wet and dry surfaces, and for both test organisms, with extracellular ATP levels present in the culture growth medium having little effect on the ATP signal detected. Hygiene swabbing of wet surfaces using direct streaking gave the best results and was superior to the ATP test in the determination of bacterial contaminants. Hygiene swabbing of dry surfaces, especially if using a diluent in the recovery stage, gave the poorest results, which were several log factors inferior to those of the ATP system and varied between test organisms.

The results obtained for wet inoculated surfaces could, theoretically, explain why in some circumstances it might be possible to recover microorganisms from a surface which an ATP test indicates is clean (9). This would require no food debris to be present coupled with an optimum swabbing recovery technique from a wet surface. However, this is unlikely to occur in situ. Microorganisms are rarely present in the form of a pure culture, and product residue is likely to be present as the major component of an unclean surface. Given increasing concerns over cross-contamination and the low minimum infective dose of pathogens such as E. coli O157, detecting low levels of contamination will become increasingly important. While future developments in swab wetting agents might improve the recovery of dry microbial bioburden, enhancing the detection limits of ATP testing is currently the subject of research by at least two companies.

The difference in results obtained using hygiene swabbing could be due to differences in bacterial adherence, or a loss in viability during drying. Studies on Listeria spp. indicate that drying quickly results in loss of viability (10). However, studies on other pathogens indicate these may survive extended periods in dry product residues on food preparation surfaces (11,12)(11,12) and it is important that monitoring methods should be capable of detecting this residue. The difference in recovery of EE. coli and S. aureus from dry surfaces could be due to the ability of the latter to tolerate dry conditions (11).

Factors which will influence the minimum detection limits of both methods will include the ability of the swab to remove bioburden from the surface, which will, to a certain extent, be influenced by the swabbing procedure used. The successful removal of the bioburden from the surface will be the result of using an effective swab wetting agent which will assist in bioburden removal and pick-up. In hygiene swabbing, reliable plate counts will only be obtained if the recovered microorganisms have been effectively released from the swab bud, either
directly onto an agar surface, or into a suitable recovery diluent which can subsequently be used for the preparation of pour or spread plates. Growth of recovered microorganisms will be influenced by the composition of the growth media used and the time and temperature of incubation. Specific to the ATP bioluminescence assay used is the requirement that the enzyme system is active throughout the reported shelf-life of the test system, and that the photomultiplier tube can detect and measure the light produced from the buffer into which the ATP is released.

Surface cleanliness testing is no longer the preserve of dedicated quality control departments and should be as simple and reproducible as possible. The data showed that the ATP bioluminescence test system had better reproducibility than traditional microbiological swabbing and this confirms previous findings (1). Recovery rates for hygiene swabbing were typical of those found in previous studies (12). It is recognized, however, that problems could occur in determining the level of microbial contamination on surfaces using hygiene swabbing due to the presence of viable but non-culturable bacteria (13).

The present data on detectability and reproducibility for different organisms, coupled with real-time results, indicate that ATP testing should be the initial method of choice in hygiene monitoring, especially within HACCP plans. However, this should be integrated with microbiological testing using optimum recovery techniques as part of a coherent surface cleanliness monitoring system (1).

REFERENCES

1. Griffith CJ, Davidson CA, Peters AC, Fielding LM. Towards a strategic cleaning assessment programme: hygiene monitoring and ATP luminometry, an options appraisal. Food Sci Technol Today 1997;11:15-24.
2. Report of the 29th session of the Codex Committee on Food Hygiene. Appendix 2: The HACCP System and Guidelines for its Application. WHO ALINORM 97/13A.
3. Griffith CJ, Blucher A, Fleri J, Fielding LM. An evaluation of luminometry as a technique in food microbiology and a comparison of six commercially available luminometers. Food Sci Technol Today 1994;8:209-16.
4. Ehrenfeid EE. ATP Bioluminescence: real-time testing of cleaning effectiveness. Food Technol Europe 1995;Septemher/October.
5. Kyriakides AL, Costello SM, Doyle G, Easter MC, Johnson I. Rapid hygiene monitoring using ATP bioluminescence. In: Stanley PE, Kricka LJ, editors. Bioluminescence and chemiluminescence: current status. Chichester: Wiley 1991;519-22.
6. Seeger K, Griffiths MW. Adenosine triphosphate bioluminescence for hygiene monitoring in health care institutions. J Food Prot 1994;57:509-12.
7. Bautista DA, McIntyre I, Laleye L, Griffiths MW. The application of ATP bioluminescence for the assessment of milk quality and factory hygiene. J Rapid Meth Auto Microbiol 1996;1:179-93.
8. Poulis JA, dePijper M, Mossel DAA, Dekkers PPhA. Assessment of cleaning and disinfection in the food industry with the rapid ATP-bioluminescence technique combined with the tissue fluid

contamination test and a conventional microbiological method. Int J Food Microbiol 1993;20:109-16.
9. Griffiths MW. The role of ATP bioluminescence in the food industry: new light on old problems. Food Tech 1996;6:62-71.
10. Cox LJ, Kleiss T, Cordier JL, Cordellana C, Konkel P, Pedrazzini C, Beumer R, Siebenga A. Listeria spp. in food processing, nonfood and domestic environments. Food Microbiol 1989;6:49-61.
11. Scott E, Bloomfield SF. The survival of microbial contamination
on worksurfaces and cloths and its transfer via cloths, hands, and utensils. Environ Health 1990;February:34-7.
12. Humphrey T, Mason M, Martin K. The isolation of Campylobacter jejuni from contaminated surfaces and its survival in diluents. Int J Food Microbiol 1995;26:295-303.
13. Bovill RA, Mackey BM. Resuscitation of ‘non-culturable’ cells from aged cultures of Campylobacter jejuni. Microbiology 1997;143:1575-81.