Abstract
Objective
To assess the effectiveness of cleaning protocols and support infection prevention efforts, we instituted ATP bioluminometer monitoring at the University of Minnesota Veterinary Medical Center. Our objective with this serial cross-sectional study was to summarize our ATP bioluminescence reading from animal and human contact areas and the impact of seasonality.
Methods
From April 2020 through November 2023, swabs were collected from common contact surfaces at the hospital. Using the 3M Clean-Trace ATP Luminometer, relative light units (RLUs) served as a surrogate for an assessment of hygiene. Relative light unit values were compared by location, season, and “animal contact” or “human contact” surfaces. A mixed model compared RLU levels across different locations and dates.
Results
ATP readings varied across locations, ranging from 5 to 301,158 RLUs, with a mean of 1,441.14 (± 8,951.79), a median of 494, and an IQR of 1,138. Most readings were below 1,000 RLUs (67%), and 37% were below 300 RLUs. Animal contact areas had lower RLU readings compared to human contact areas. The mixed model identified statistically significant variable RLU values by location but not by season.
Conclusions
We observed a wide variation in median RLU values across the sampled locations. This is to be expected since hospital environments are dynamic, with varied animal and human interactions throughout the day as well as changing staffing patterns and patient volumes across different seasons.
Clinical Relevance
Maintaining high hygiene standards is crucial for patient well-being and reducing the risk of healthcare-associated, zoonotic, and antimicrobial-resistant infections. The use of the ATP bioluminometer is one tool to reduce healthcare-associated infections and support better patient outcomes.
Healthcare-associated infections (HAIs) are a recognized problem in both human and veterinary care settings. It is estimated that 3.2% of hospitalized human patients may develop an HAI.1 As a result of this concern, a number of measures have been implemented to protect patient safety, including surveillance for antimicrobial-resistant organisms, implementation of standard precautions, improved hand hygiene, monitoring of environmental contamination, and disinfection of patient and staff contact surfaces and equipment. The recent detection of a cluster of carbapenem-resistant Enterobacterales cases among companion animals in a tertiary veterinary clinic highlights this concern, especially with the potential environmental contamination and spread to other patients.2
The implementation of cleaning and disinfection protocols is a key effort to reduce HAIs in human healthcare settings. Monitoring the effectiveness of cleaning practices is one recommendation to ensure appropriate environmental cleaning and a reduction of the risk of HAIs.3 To determine whether the environment is sufficiently cleaned, visual inspection is often the primary method used. Other assessment methods include chemical or microbiological assessments, such as using ATP bioluminescence or aerobic colony counts.3
To monitor the effectiveness of the Veterinary Medical Center (VMC) cleaning and disinfection practices, we began routine environmental hygiene assessments using an ATP bioluminometer in our hospital. Although ATP assessment is used in human hospitals and food establishments, its adoption in veterinary hospitals is limited. Our objectives were to summarize our initial findings and specifically describe the application of ATP bioluminescence readings from animal and human contact areas and potential changes due to season.
Methods
Environmental assessments
The Infection Control Surveillance program at the VMC includes periodic hygiene assessments. This became relevant during the COVID-19 pandemic to address hygiene concerns. In 2020, 19 areas that were commonly used for patient care or were common hand-contact surfaces were selected for routine hygiene assessment utilizing the Clean-Trace ATP Luminometer (3M). These sites were selected based on the risk assessment of the infection prevention team. Sampling locations included 4 examination tables, 3 breakroom tables, 3 controlled drug machines (CDMs), 3 computer keyboards, 2 door handles, 1 urine sample handling station, 1 telephone, an elevator button, and a button on a patient scale. For each sampling location, bioluminometer swabs were collected weekly from late April 2020 through mid-June 2020, biweekly from July 2020 through November 2021, and monthly from December 2021 forward.
Samples were collected, activated, and measured as recommended in the 3M sampling guide. First, ATP swabs were removed from refrigeration, allowed to come to room temperature for at least 10 minutes, and labeled with the testing location. Samples were collected over a 4 X 4-inch square applying gentle downward pressure while rotating the swab, sampling the surface left to right, then top to bottom before placing the swab back in the test tube. Swabs were placed into the chemical at the bottom of the tube and shaken side to side for 5 seconds to ensure mixing of reagents prior to placing it in the Clean-Trace ATP Luminometer for measurement. The Clean-Trace ATP Luminometer measures ATP as relative light units (RLUs). For the purpose of these routine hygiene assessments, the VMC uses ≤ 1,000 RLUs as the “pass” value and ≥ 1,001 as the “fail” value. There are currently no standard uniform protocols set for thresholds that determine the level of cleanliness in veterinary settings. In the literature, a common threshold pass rate for ATP hygiene assessments in human hospitals is around 250 or 500 RLUs, although variations exist due to different benchmarks set by various tools.3–5 We established our cutoffs values based on historical median and mean RLU values collected longitudinally overtime.
The Infection Control Surveillance program at the VMC also conducts hygiene assessments of ICU cages used to house known or suspected infectious patients. As these cages are swabbed after terminal cleaning, the VMC uses ≤ 300 RLU as the “pass” value and ≥ 301 as the “fail” value. These values are based on historical applications and values used in human hospital settings.3,4,6,7
Statistical analysis
This serial cross-sectional study analyzed ATP bioluminometer values collected at regular intervals. The data cleaning process involved categorizing locations as either “animal contact” or “human contact” to facilitate the assessment of differences. Sample collection dates were also converted to facilitate the assessment of seasonal trends: we defined the winter of a year as December of the previous year, along with January and February of the current year. Spring was defined as March, April, and May of the current year. Summer was defined as June, July, and August of the current year, and fall was defined as September, October, and November of the current year.
Using R statistical software (version 4.3.3; R Foundation for Statistical Computing)8 and Excel (Microsoft Corp), we conducted an initial descriptive statistical analysis summarizing RLU values across all of the locations. This included means, medians, SDs, and IQRs as well as calculating the proportions of readings that were deemed to be clean (RLUs < 1,000). Box plots were created to visualize the distribution of RLU levels by location, season, and contact type (animal vs human). Animal contact areas were specifically defined as the examination tables. All other areas were categorized as human contact areas, which reflected areas touched by human hands (ie, computer keyboards, elevator buttons, and pharmacy dispensing devices [CDMs]).
To reduce the effects of skewed RLU distributions and the impact of high outlier values, which could obscure underlying patterns, we applied a logarithmic transformation to the RLU data, converting RLU levels to a natural log scale. Afterward, we conducted pairwise comparisons using the Tukey honestly significant difference adjustment method to pinpoint locations with statistically significant differences in RLU levels, employing a compact letter display and significance level of 0.05.
We also compared RLU levels across different locations and dates using a mixed model. This model incorporated fixed effects for location and season and random effects for the date of sample collection to account for both fixed and random variations in RLU levels. Differences between locations and seasons were assessed using the Tukey adjustment method.
Results
From April 2020 through November 2023, 1,193 environmental samples were collected. The ATP readings varied, with a range from 5 to 301,158 RLUs. The mean reading was 1,441.14 (± 8,951.79), with a median of 494 and an IQR of 1,138. Table 1 provides a summary of values across all 19 site locations.
Site monitoring by relative light unit (RLU) at the Veterinary Medical Center (April 2020 through November 2023).
| ATP value (RLUs) per location | Obs | Mean ± SD | Median (IQR) | 25th percentile | 75th percentile | Range | No. clean (< 1,000) | No. clean (< 100) |
|---|---|---|---|---|---|---|---|---|
| Break area: break room | 59 | 1,161 (± 1,413) | 669 (944) | 284.5 | 1,202.5 | 24–7,391 | 38 (64.4%) | 4 (6.8%) |
| Break area: FTR | 59 | 1,050 (± 1,381) | 668 (1,066) | 274.5 | 1,331.5 | 31–8,547 | 39 (66.1%) | 9 (15.3%) |
| Break area: VPM | 59 | 1,278 (± 1,439) | 786 (1,213) | 311.5 | 1,522.5 | 31–7,371 | 34 (57.6%) | 4 (6.8%) |
| Button: elevator | 59 | 251 (± 326) | 132 (275) | 45.5 | 315 | 14–1,527 | 57 (96.6%) | 24 (40.7%) |
| Button: scale in lobby | 59 | 1,838 (± 2,753) | 523 (1,805) | 264.5 | 1,894 | 29–13,873 | 34 (57.6%) | 7 (11.9%) |
| CDM: ER | 67 | 3,298 (± 4,921 | 1,326 (3,683) | 388.5 | 3,903 | 15–22,458 | 32 (47.8%) | 5 (7.5%) |
| CDM: ICU | 67 | 1,861 (± 1,753) | 1,379 (1,420) | 857 | 2,243 | 34–10,674 | 23 (34.3%) | 2 (3.0%) |
| CDM: TR | 67 | 1,129 (± 1,784) | 468 (1,018) | 223 | 1,198.5 | 17–11,502 | 46 (68.7%) | 8 (11.9%) |
| Door hallway ICU/exercise yard | 59 | 1,931 (± 4,041) | 678 (1365) | 258.5 | 1,530 | 10–2,5621 | 34 (57.6%) | 4 (6.8%) |
| Main entry door handle | 59 | 1,065 (± 1,509) | 634 (927) | 330 | 1,228 | 16–9,243 | 40 (67.8%) | 5 (8.5%) |
| Examination table ER by cages | 67 | 5,671 (± 36,769) | 266 (1,007) | 119 | 1,099.5 | 6–301,158 | 48 (71.6%) | 15 (22.4%) |
| Examination table ICU (by cage) | 67 | 296 (± 538) | 100 (206) | 56 | 249 | 7–2,437 | 61 (91%) | 33 (49.3%) |
| Examination table TR (by U/S) | 67 | 958 (± 1,527) | 300 (999) | 129 | 1,103 | 5–8,453 | 49 (73.1%) | 13 (19.4%) |
| Examination table UC (by hallway door) | 67 | 747 (± 1,226) | 344 (884) | 122 | 1,001 | 13–8,952 | 50 (74.6%) | 15 (22.4%) |
| ICU computer (by lab) | 67 | 1,257 (± 1,181) | 710 (1,460) | 390 | 1,832.5 | 36–5,046 | 37 (55.2%) | 4 (6.0%) |
| Shared computer by CP | 59 | 815 (± 922) | 529 (782) | 241.5 | 1,002.5 | 28–5,287 | 44 (74.6%) | 7 (11.9%) |
| Shared computer by ER | 59 | 958 (± 1,044) | 550 (1,175) | 196 | 1,344 | 41–4,448 | 40 (67.8%) | 6 (10.2%) |
| Shared phone by ER | 59 | 856 (± 1,079) | 487 (862) | 183 | 1,029.5 | 18–5,048 | 42 (71.2%) | 4 (6.8%) |
| TR urine sample station | 67 | 580 (± 795) | 259 (715) | 89 | 699.5 | 10–3,844 | 54 (80.6%) | 18 (2.7%) |
CDM = Controlled drug machine. CP = Clinical pathology. ER = Emergency room. FTR = Float technician room. Obs = Observation. TR = Treatment room. UC = Urgent care. U/S = Ultrasound. VPM = Veterinary Population Medicine department.
Break area: break room refers to the table in the break space on the second floor. Door hallway ICU/exercise yard refers to the door handle in the hallway between the ICU and exercise yard. Shared computer by CP refers to the shared computer near the CP laboratory receiving.
There was a wide variation in the median RLU values across locations and years (Figure 1). Using the threshold of 1,000 RLUs, the overall results suggest that cleaning efforts were consistent and generally below the threshold. This was evident by year and season (Figure 2). Relative light unit values below 300 were recorded in 37.4% of samples.
Distribution of relative light unit readings across sites (April 2020 through November 2023) on a logarithmic scale. CDM = Controlled drug machine. Clin path = Clinical pathology. ER = Emergency room. TX RM = Treatment room. UC = Urgent care. u/s = Ultrasound. VPM = Veterinary Population Medicine department.
Citation: American Journal of Veterinary Research 86, 6; 10.2460/ajvr.24.09.0278
Box plot of ATP readings by season and year, log scale.
Citation: American Journal of Veterinary Research 86, 6; 10.2460/ajvr.24.09.0278
In the animal contact areas, the RLU readings were lower overall, with 75% of the RLU values falling below the threshold. In areas of human contact, there was a wider range of RLU values. The examination table in the emergency room recorded the highest total RLU values, significantly influenced by an outlier with an RLU value of 301,158. The next highest values were observed at the CDM in the emergency room, followed by those at the CDM in the ICU.
In the pairwise comparisons between locations, we observed statistically significant differences in mean RLU values across multiple locations, but there did not appear to be any observable patterns in differences.
Comparing RLU results across seasons demonstrated considerable variation in RLU values, with geometric mean values ranging from 498 (fall) to 555 (spring; Figure 2). In the mixed model accounting for location and season, we found no statistically significant difference between seasons, but statistically significant differences remained between some locations.
Discussion
There is an ongoing discussion about monitoring the veterinary environment to reduce bacterial contamination.9 Early attempts have focused on microbiologic surveillance of the environment; however, the CDC does not recommend routine environmental culturing.10 This was because rates of HAIs were not associated with microbial contamination of air or environmental surfaces, and meaningful permissible standards of microbial contamination of environmental surfaces did not exist.10 Yet, there is value in timely, less expensive ways to evaluate environmental contamination. The use of ATP bioluminescence offers a valuable tool to assess and provide feedback on veterinary hospital infection control practices.11 Our routine incorporation of ATP bioluminescence identified areas of concern and reminded staff to perform regular cleaning and disinfection to reduce biologic contamination loads in busy clinic areas. With our assessment of common contact areas for patients and staff, we did observe higher ATP bioluminescence values on human contact surfaces versus animal contact surfaces. Even though not statistically significant, this observation was expected based on previous unpublished assessments.
We also observed a wide variation in median RLU values across the sampled locations. We were expecting to observe some seasonality in our RLU observations, especially during the busy summer season. This is to be expected since hospital environments are dynamic, with varied animal and human interactions throughout the day as well as changing staffing patterns and patient volumes across different seasons. This can be compounded by the buildup of microbiologic contamination and growth and the development of mechanisms that impair cleaning (ie, biofilms). This variability is affected by the use, cleaning, efficiency of disinfectants, and the experience and time availability of staff. An example is the pharmacy dispensing devices (eg, CDMs) commonly used by staff. These dispensing machines are busy, central areas of the hospital and contacted by human hands from a variety of staff, faculty, and students. We observed that the CDMs often had consistently higher levels of contamination compared to the animal contact surfaces. This served as a reminder for staff to practice regular hand hygiene and to implement routine cleaning procedures to reduce contamination of these high-hand-contact surfaces. Similarly, we had demonstrated the contamination of computer keyboards and the value of routine cleaning to reduce staphylococcal contamination.11 The ATP bioluminescence provided valued feedback to staff on areas of concern and evaluation of regular and routine cleaning and disinfection of common hospital surfaces. Additionally, our infection prevention staff have utilized the bioluminometer on easily contaminated equipment, such as endoscopes and ultrasounds. Endoscopes are common and important diagnostic tools but are difficult to clean.
In addition, the ATP surveillance provided an opportunity to review and update our cleaning and disinfection protocols. This includes specific protocols for cleaning after infectious patients in the ICU, deep cleaning after handling infectious patients, and postcleaning assessment of cages and other equipment. In addition, pocket-sized educational guides are annually provided to each student entering clinic rotations, new house officers, and staff. These educational materials are reinforced with posters, clinic newsletters, and educational rounds. One of the limitations of this study is the lack of standardized RLU levels in the veterinary setting. We determined our RLU levels historically over time to establish our hospital-specific cutoffs based on median and mean values. We recognize that these values will vary by bioluminometer, surface types, and cleaning protocols. Occasionally, very high values have been documented, like the extreme value of 301,158 RLUs. The reason for this value remains unclear, raising questions about whether it was due to a measurement error, an extraordinarily unsanitary condition, or an unusual contamination event. Having a framework to investigate unique circumstances that lead to such extreme readings is crucial for refining and updating infection control policies. Understanding these anomalies helps accurately identify and focus on high-risk areas or activities, which in turn can result in targeted measures for maintaining optimal hygiene standards. Infection prevention staff also utilized the ATP bioluminescence to assess hygiene after a suspected or known infectious patient was discharged. This served as a reminder for thorough cleaning and raised awareness for staff to take added precautions before placing another patient in that cage.
It is plausible that staff may have become aware of the sampling protocols and timing and enhanced cleaning when they anticipated that sampling would occur. However, infection prevention staff did not announce their sampling plans. The sampling results were shared with staff and hospital management at monthly infection prevention and control meetings.
In the future, the utilization of bioluminescence methods at other veterinary facilities would support the standardization of hygiene thresholds. Ultimately, these methods need to be paired with strong surveillance systems that document HAIs. Ideally, these methods, bundled with other infection prevention tools, can document reduced HAIs and support better patient outcomes. Maintaining high hygiene standards is crucial for patient well-being and reducing the risk of healthcare-associated, zoonotic, and antimicrobial-resistant infections.12
Acknowledgments
None reported.
Disclosures
The authors have nothing to disclose. No AI-assisted technologies were used in the composition of this manuscript.
Funding
Partial funding support for antibiotic stewardship and multidrug-resistant infection surveillance was provided by CDC One Health Collaboration to Combat Antimicrobial-Resistant Infections Grant H1201-3000068545.
ORCID
O. Akinyede https://orcid.org/0000-0003-3439-4178
M. V. Boyd https://orcid.org/0009-0002-9460-6224
K. E. Frerichs https://orcid.org/0009-0001-8395-3258
A. Rendahl https://orcid.org/0000-0001-5434-3592
J. B. Bender https://orcid.org/0000-0003-4645-9936
References
- 1.↑
Magill SS, O’Leary E, Janelle SJ, et al. Changes in prevalence of health care–associated infections in U.S. hospitals. N Engl J Med. 2018;379(18):1732–1744. doi:10.1056/NEJMoa1801550
- 2.↑
Cole SD, Peak L, Tyson GH, Reimschuessel R, Ceric O, Rankin SC. New Delhi metallo-β-lactamase-5–Producing Escherichia coli in companion animals, United States. Emerg Infect Dis. 2020;26(2):381–383. doi:10.3201/eid2602.191221
- 3.↑
Sherlock O, O’Connell N, Creamer E, Humphreys H. Is it really clean? An evaluation of the efficacy of four methods for determining hospital cleanliness. J Hosp Infect. 2009;72(2):140–146. doi:10.1016/j.jhin.2009.02.013
- 4.↑
Nante N, Ceriale E, Messina G, Lenzi D, Manzi P. Effectiveness of ATP bioluminescence to assess hospital cleaning: a review. J Prev Med Hyg. 2017;58(2):E177–E183.
- 5.↑
Neogen Clean-Trace ATP Monitoring System. Accessed March 10, 2025. https://www.neogen.com/clean-trace-healthcare/
- 6.↑
Whiteley GS, Glasbey TO, Fahey PP. Using a simplified ATP algorithm to improve data reliability and improve cleanliness standards for surface and medical device hygiene. Infect Dis Health. 2022;27(1):3–9. doi:10.1016/j.idh.2021.07.001
- 7.↑
Deshpande A, Dunn AN, Fox J, et al. Monitoring the effectiveness of daily cleaning practices in an intensive care unit (ICU) setting using an adenosine triphosphate (ATP) bioluminescence assay. Am J Infect Control. 2020;48(7):757–760. doi:10.1016/j.ajic.2019.11.031
- 8.↑
R: a language and environment for statistical computing. Version 4.3.3. R Foundation for Statistical Computing. Accessed July 15, 2024. https://www.R-project.org
- 9.↑
Allerton F, Weese S. CON: environmental microbiological surveillance does not support infection control in veterinary hospitals. JAC Antimicrob Resist. 2024;6(4):dlae114. doi:10.1093/jacamr/dlae114
- 10.↑
Sehulster L, Chinn RYW; CDC; HICPAC. Guidelines for environmental infection control in health-care facilities: recommendations of CDC and the Healthcare Infection Control Practices Advisory Committee (HICPAC). MMWR Recomm Rep. 2003;52(RR-10):1–42.
- 11.↑
Bender JB, Schiffman E, Hiber L, Gerads L, Olsen K. Recovery of staphylococci from computer keyboards in a veterinary medical centre and the effect of routine cleaning. Vet Rec. 2012;170(16):414. doi:10.1136/vr.100508
- 12.↑
O’Neill J. Antimicrobial resistance: tackling a crisis for the health and wealth of nations. AMR Review. December 2014. Accessed July 30, 2024. https://amr-review.org/sites/default/files/AMR%20Review%20Paper%20-%20Tackling%20a%20crisis%20for%20the%20health%20and%20wealth%20of%20nations_1.pdf
