The assessment of a hygienic state or cleanliness of contact surfaces has significant implications for food and medical industries seeking to monitor sanitation and exert improved control over a host of operations affecting human health. Methods used to make such assessments commonly involve visual inspections, standard microbial plating practices, and the application of ATP-based assays. Visual methods for inspection of hygienic states are inherently subjective and limited in efficacy by the accuracy of human senses, the degree of task-specific work experience, and various sources of human bias. Standard microbial swabbing and plating techniques are limited in that they require hours or even days of incubation to generate results, with such steps as enrichment and colony outgrowth resulting in delays that are often incompatible with manufacturing or usage schedules. Rapid in conduct and considered more objective in operation than visual or tactile inspection techniques, swabbing surfaces using ATP-based assessments are relied on as routine, even standard, methods of hygienic assessment alone or in complement with microbial and visual inspection methods. Still, current ATP methods remain indirect methods of total hygiene assessment and have limitations that must be understood and considered if such methods are to be applied judiciously, especially under increasingly strict demands for the verification of hygiene state. Here, we present current methods of ATP-based bioluminescence assays and describe the limitations of such methods when applied to general food manufacturing or health care facilities.
The assessment of the hygienic state or cleanliness of contact surfaces in the food and medical industries is a major task requiring significant diligence, time, and expense. Traditionally, visual and tactile inspections, complemented by standard microbial swabbing or plating practices, were the sole means of making such assessments. However, beginning in the 1990s, bioluminescence assays based on ATP have been increasingly used in food industries and more recently in medical industries as key methods of hygienic assessment alone or in complement with conventional microbial and inspection methods. Initially adopted in the 1960s both by scientists at the National Aeronautics and Space Administration (NASA) looking to determine levels of biological material (27, 62, 63, 70) and marine biologists looking at ocean biomass (51), the basis of the ATP bioluminescence assay has been studied over several decades (4, 33, 37, 74–76, 112, 114, 115). ATP bioluminescence is centered on the firefly luciferin-luciferase enzyme reaction, in which the luciferase enzyme is exposed to ATP and oxygen, converting luciferin to oxyluciferin with the emission of light. This resulting light can be detected using a luminometer, and its intensity corresponds to the amount of ATP present in the sample. A nearly ubiquitous energy-bearing molecule in biology, the resulting light intensity is interpreted as a measure of the degree of biological surface contamination. Rapid in conduct and considered more objective in operation, ATP-based assessments measure ATP levels of a surface, liquid or substance, that are indicative of the presence of biological matter, including food residues, bodily fluids, and microorganisms.
ATP-BASED ASSAYS IN APPLICATION
ATP-based assessments were first used to test concentrations of ATP in food products in the 1970s (99) and on surfaces in the 1980s (73, 100, 114); however, reagents were still expensive and testing was still lengthy (114). Over time, the introduction of rapid handheld luminometers and less expensive and better reagents that could go longer before decay postreaction improved the time, reproducibility, and cost efficiency. By the 1990s, ATP bioluminescence-based assays became more commonplace as a key component in response to regulatory and industry demands to better assure surface sanitation in the food industry (13, 15, 40, 42, 43, 89). ATP-based assessments allowed rapid assessment of hygienic state or cleanliness of contact surfaces, including as a means of process verification in the greater context of food manufacturing systems based on hazard analysis and critical control point (32). Since the early 1990s, ATP bioluminescence assays have gained substantial value in the food industry for dairy (13, 15, 43), meat (14, 89), brewing (103), and other processed food production as a means of hygiene monitoring. Currently, ATP-based assays are typically conducted using a specialized swab applied to the desired test area and read using a handheld luminometer, which provides results in relative light units (RLUs). Surfaces such as stainless steel, Teflon, tiles, plastics and polymers, gaskets, and milk processing machinery are commonly tested for various nonvisual soils, such as processing residues, microbiological matter, allergens, and other bioburdens (89, 124). The rapid assessment of surface contamination by microorganisms, both culturable and nonculturable; residual allergens; or biological material that can support growth of microorganisms is a key goal for manufacturing industries.
In more recent years, medical industries have adopted ATP bioluminescence testing, first to test the cleanliness of various surfaces in hospital kitchens or canteens (10, 39, 96), testing sinks, cutting boards (60), meat cutters, and knives for biological matter, especially where pathogen control was critical. Starting in the early 2000s, several initiatives were launched in the United Kingdom to combat nosocomial or hospital-acquired infections within the National Health System, which saw the introduction over the next decade of ATP bioluminescence hygiene testing of surgical equipment, intensive care areas, surgical operating theaters, nurse stations, and patient areas, especially high-touch objects such as bed controls, phones, call buttons, light switches, intravenous (i.v.) poles, doorknobs, and other machine buttons for microbial contamination or contamination from bodily fluids and other biological soils (10, 29, 39). However, although some surfaces tend to be hygienically designed, many surfaces in either food manufacturing or hospital settings are often porous or otherwise improperly designed with regard to sanitation (98). In addition, because immediate recontamination is all but guaranteed in most food manufacturing and medical areas due to constant use, ATP bioluminescence is often used more as an indicator of effective cleaning and sanitization or to validate new cleaning methods (65).
CURRENT KNOWLEDGE OF ADENYLATE METABOLISM
The underlying basis of the ATP-based bioluminescence assay is rooted in the understanding that all living organisms generate and maintain a pool of ATP, which in general is used as an energy-imparting molecule for a host of metabolic reactions (9, 26, 67). It is acknowledged that there are exceptions, such as bacterial spores that have shown to have little to no ATP levels (approximately 10−21 mol ATP per spore versus 10−17 mol ATP per bacterium (58, 120)) and thus may not display sufficient ATP for ATP-based hygiene assays (97). However, studies by NASA scientists on the degradation of both intracellular and extracellular ATP on the surface of spacecraft have contradicted the typically accepted idea of immediate degradation of ATP at time of cell death and indicated that free ATP from lysed cells may remain present longer than previously thought (94, 120). The ability of some residual ATP to remain present, as well as trends demonstrating that most ATP decays into ADP and AMP, are supported by subsequent work on adenylate homologue profiles over the course of significant metabolic changes in various organisms (105, 106).
A basis of ATP bioluminescence carries assumptions of reasonably high and stable ATP concentrations from a cell, tissue, or biological material so as to allow a correlation between the ATP concentration levels and the bioburden on a particular surface. Although some studies have demonstrated that the concentrations of ATP within a given organism are reasonably constant under a given condition (42, 111, 112), different bacterial species have been shown to have as much as a 30-fold difference among ATP concentrations per cell of bacterial species, ranging from 0.28 to 8.90 μg of ATP per bacterial cell (27), with differences in cellular mass serving as a significant source of variation. ATP concentrations in yeast and mammalian cells are around 100-fold higher than in bacterial cells (99). On the average, there are 10−15 g of ATP per bacterial cell compared with 10−12 g of ATP per yeast or mammalian cell (112) or, as reported in moles, an average of 10−18 mol of ATP per bacterial cell and 10−15 mol of ATP per mammalian cell (67). It has also been shown that the age of the cells, stage in the growth cycle (61, 106), nutritional status, oxygen levels (72), pH (6, 44, 104), temperature (80), and even specific organism can influence ATP content (9, 44, 104, 111, 131) through ATP synthesis, utilization, or decay.
LIMITATIONS OF ATP-BASED ASSAYS
Although ATP bioluminescence has many uses and benefits, it is important to recognize and consider the limitations of current ATP-based methods. One of the major limitations is the inability to directly compare RLU values or results from different methods or ATP test systems. Brands of luminometers can read differently when compared, even differing sometimes from model to model of same manufacturer (5, 79). RLUs cannot be used in direct cross-comparison; instead, the luminometer values must be converted back to ATP concentration from a standard curve or some other internal comparison. Even then, comparisons of different luminometers (5, 86, 95, 127, 128) have been shown to have differing minimum detection limits, differing levels of quenching or enhancement effects from disinfectants, or differing abilities to detect ATP from various environments. Furthermore, various surfaces swab differently; smooth surfaces may be more easily swabbed than rough ones (19), and various surfaces can be difficult to accurately swab because of surface porosity (88) or accessibility. Differing brands of luminometers have variations in swab tips or swabbing methods, which could also contribute to these brand-to-brand effects (95). Collectively, these variables place a responsibility on the user to carefully select the best luminometer for the application.
Although ATP bioluminescence methods are valuable tools for real-time detection, bioluminescence methods are generally less suited to applications in which an especially low bioburden is present or spore contamination is the specific target for assessment. Furthermore, ATP-based assays cannot differentiate ATP-generating microorganisms from organic debris containing ATP and thus should not be considered replacements for routine, microbiological testing as a means of verification. Similarly, RLUs do not correspond directly with CFUs, although the trends are similar (i.e., high CFUs usually have high RLUs, but not necessarily vice versa).
Although many of these problems are minimized or eliminated with proper internal controls, most assay methods do not seem to routinely incorporate sufficient internal controls or when developing or adjusting testing methods, including controls to verify upper and lower limits of the luminometer for each protocol, which are necessary for proper use. This partly results from the current lack of a generally accepted standard for internal controls, because the bioluminescence reaction of a luminometer does not process and detect purified stock ATP applied directly onto a swab in the same manner as free ATP found on a swabbed surface or even ATP from bacteria swabbed from a surface. It has been shown that there is some difference in certain studies between detection of gram-positive and gram-negative bacteria, theoretically because of difficultly lysing of the cell in the reaction (118). One recommendation established upper limits based on swabs before cleaning and sanitation and established lower limits from swabs after proper cleaning and sanitation (48) for every tested environment, with the additional recommendation of completion of 10 consecutive cleaning and sanitation cycles to establish an obtainable lower limit (50).
Lastly, specific chemicals, sanitizers, disinfectants, or even residual chemicals found in cleaning cloths can quench results or create errors if residual chemicals come in contact with the proprietary ATP test reagents. Chemicals shown to exhibit this quenching effect include trisodium phosphate, lactic acid, hydrogen peroxide, triclosan, and sodium hypochlorite (22, 38, 59, 119). Although beyond the scope of this article, additional factors within luminometer design, such as system software or selection of photodiode versus photomultiplier tubes, may manifest variations in performance parameters including test sensitivity, initial and maintenance costs, and unit longevity.
BENCHMARKS IN ATP BIOLUMINESCENCE
Another major limiting factor of ATP-based assays is the lack of a recognized standard of cleanliness, both in method of testing and in standards for pass-or-fail indicators. Although it seems to be generally accepted that an area 10 by 10 cm (or 100 cm2) is appropriate for swabbing (18, 32), other test sizes are often seen (Table 1), especially with irregularly shaped objects, smaller buttons, and touchpads.
A benchmark of what is considered clean varies with the industry and instrument and can range anywhere from <100 to <500 RLU (Table 1). The food industry seems to generally adhere to more stringent levels of cleanliness, partly because surfaces can be more contained when not in use or even during use, whereas the medical industry tends to accept higher pass levels, with the exception of surgical equipment. ATP testing methods for medical settings proposed by Griffith et al. (39), seem to be the most cited reference standard with a target level of cleanliness of <500 RLU, although these recommendations were later modified to a target level of <250 RLU with the proviso that this target RLU was specifically for the luminometer used in the study (65). This RLU-based circumstance was noted by Boyce et al. (18) when they observed lower median RLU values than those previously reported (39, 65) using the same brand of luminometer. For all these reasons, the establishment of a user-based, validated protocol for establishing a pass-or-fail benchmark may be more appropriate than having a universal benchmark RLU benchmark.
In addition, differing assay protocols can potentially minimize or maximize the results, leading Ho et al. (49) to recommend the use of RLU per square centimeter as the standard unit, similar to units used in microbial swabbing. As they discussed (49), and as shown more fully in Table 1, a lower RLU benchmark will seem more stringent, but differing the surface area tested can make a higher RLU benchmark more stringent (i.e., 500 RLU from 100 cm2 is 5 RLU/cm2 versus 100 RLU from 10 cm2 is 10 RLU/cm2). Accordingly, common brands of luminometers have different settings for pass or fail and even have various settings for a machine for different methods or testing conditions. Recommended pass or clean RLUs given for common brands of luminometers are as follows: The Clean-Trace/Bio trace (3M, St. Paul, MN) recommends cleanliness with ≤250 RLU for environmental surfaces with 100-cm2 test areas (1), with ≤200 RLU for surgical endoscopes (before final sterilization) (2), and with ≤150 RLU for surgical equipment (3). The AccuPoint Advanced ATP Reader (Neogen, Lansing, MI) also suggests a test area of 10 by 10 cm with preprogrammed default threshold settings of ≤149 RLU as pass or clean and ≥300 RLU as fail (84). The Lumitester PD-30 (Kikkoman, Tokyo, Japan) has three preprogrammed modes with pass-or-fail limits set as pass or clean with RLU ≤ 1,500, 500, and 200 and corresponding failure with RLU > 3,000, 1,000, and 400 (57). The SystemSURE Plus (Hygiena, Camarillo, CA) defaults to 10 and 30 RLU for pass and fail thresholds, respectively (54).
ALTERNATIVE ADENYLATE ASSAYS
Although the sensitivity of the traditional ATP-based bioluminescence assays is generally accepted for many assessment applications, greater sensitivity or additional consideration may be advised, including the need to detect low microbial contamination levels or testing for the presence of especially immunogenic ATP-containing allergens (124). To that end, various alternative methods of adenylate testing have been proposed. Cumulative adenylate levels (total adenylate [AXP]) tend to undergo less of an overall concentration change in comparison to ATP concentrations alone as dictated by growth or environmental changes (113); such changes would most likely influence the accuracy and precision of bioluminescence testing.
Squirrell and Murphy (109, 110) proposed the use of the intracellular enzyme adenylate kinase as a bacterial cell marker in place of ATP, as depicted in Figure 1. Using ADP as the substrate to drive the reaction in the direction of ATP synthesis, the standard firefly luciferase-luciferin reaction was then used to detect the ATP generated from adenylate kinase activity. In theory, this would make 40 to 50 times as much ATP available for bioluminescence. Further experimentation demonstrated a lower limit detection of 101 to 102 cells versus 104 cells with standard methodology (109, 110). In 2000, Corbitt et al. (30) compared the Squirrell and Murphy method with standard ATP bioluminescence and found that adenylate kinase bioluminescence was more sensitive than traditional ATP bioluminescence for certain foods, especially meat tissues because of the high levels of adenylate kinase in muscle tissue. However, no advantage in sensitivity was seen in testing of pure bacterial cultures with the Squirrell and Murphy adenylate kinase bioluminescence method over standard ATP bioluminescence (30).
In 1994, Hawronskyj et al. (47, 48) used an ATP recycling system to amplify all AMP into ATP with the use of adenylate kinase and pyruvate kinase (Fig. 1), where the ATP concentration was the rate-limiting factor for adenylate kinase. This system used updated bioluminescence technology but was related to methods proposed in prior years (52, 56, 68, 108). In 1978, Holm-Hansen and Karl (52) used the luciferase-luciferin reaction to determine ATP concentrations, using adenylate kinase, pyruvate kinase, and phosphoenolpyruvate (PEP) to measure AXP levels. ATP+ADP levels were determined with just pyruvate kinase and PEP, and AMP levels were calculated from calculations of the differences (52). In 1986, Lundin et al. (68) measured AXP levels in a single aliquot using adenylate kinase, pyruvate kinase, and PEP to convert all AMP and ADP into ATP and measured with luciferase reaction, although cytidine triphosphate was also used to help the reaction to an endpoint.
More recently, another ATP recycling method was proposed that used ATP+AMP monitoring in which the AMP and pyrophosphate produced from ATP by firefly luciferase were converted back into ATP by pyruvate orthophosphate dikinase (PPDK) (91, 92) (see Fig. 1). The signal strength was ATP dependent, but the system allowed long-lived RLU signals even with low numbers of swabbed ATP molecules. Later, Bakke and Suzuki (11) suggested a combination of this PPDK recycling method and pyruvate kinase recycling to allow measurement of AXP, eliminating changes from ATP degradation by measuring the AXP pool. This difference was seen recently in studies that used this method to test ATP, AMP, ADP, and AXP levels of animal carcasses postslaughter (106), test bacterial growth over extended periods (106), and assess the hygienic state in health care settings (12).
Although rapid in conduct and considered more objective in application than previous methods of hygiene assessment, ATP-based assessments are relied on as a routine method of choice. Under the ever-increasing regulatory need to use validated methods for analyses affecting health and safety, several technologies considered in this report have obtained certification under AOAC International's Performance Tested Methods Program. Yet current ATP methods remain indirect methods of total hygiene assessment and have limitations that must be understood if such methods are to be applied judiciously. In our attempt to survey current methods of ATP-based bioluminescence assays and the limitations of such methods, it has become readily apparent that the benefits of these ATP-based assays may not be yet fully realized or fully refined as a technology. A better use of positive and negative controls, standardization of units compared with areas swabbed, calibration with ATP standards, and more concrete definitions of what is clean in various industries or medical settings would allow better utilization of the full advantages of these ATP-based bioluminescence assays, including increased sensitivity and lower rates of false-negative outcomes. In addition, to accurately apply ATP-based technologies, an understanding of the reality of ATP variations over time and environment is necessary. Alternative or refined methods of ATP-based assays may advance ATP-based technologies and improve rapid surface hygiene assessments.
The authors acknowledge funding and technical assistance for this work from the Kikkoman USA R&D Laboratory (Madison, WI).