A Pilot Field Evaluation of Organic Surface Contamination in Pig Farrowing Units Using Rapid Hygiene Monitoring Methods

Abstract

Rapid and reliable detection methods are essential for routine monitoring of environmental hygiene on farms. This pilot study evaluated luminometers (LUM) and mobile flow cytometer (MFC) for assessment of surface organic contamination in farrowing units. The study was conducted on two pig farms after animal removal prior to sanitation, with sampling performed at heated pads, pen walls, and corridors. ATP measurements were carried out using three luminometers (Clean-Trace™ LM1, EnSure, and SystemSURE Plus), and residual particles were detected using a mobile flow cytometer (Cytoquant). Microbiological cultivation (TMC 36 °C) was additionally included. Significant differences in log RLU values were observed between LUM, with large effect sizes indicating a substantial influence of device type on RLU values. A high correlation was confirmed only between EnSure and SystemSURE Plus (r_s = 0.81–1.00; p < 0.05), and no relationship was confirmed between LUM and MFC (r_s = −0.49–0.77; p > 0.05). Correlations between rapid detection methods and microbiological cultivation were inconsistent. Corridors demonstrated the highest microbiological contamination, whereas MFC identified heated pads as sites with increased residual particulate contamination. The results indicate that LUM, MFC, and microbiological cultivation characterize different dimensions of environmental contamination and should therefore be interpreted as complementary rather than interchangeable methods.

1. Introduction

Environmental hygiene monitoring plays a key role in protecting animal health, welfare, and productivity in livestock production systems [1]. Maintaining cleanliness in animal housing is an integral part of biosecurity, with particular importance in facilities characterized by a high density of animals [2]. A strong emphasis on biosecurity measures can be observed not only in poultry production but also in pig farming, especially during the fattening period and, most critically, in farrowing units [3]. The aim is to minimize the risk of pathogens entering the farm and to prevent their spread in the environment [3]. An important tool in this context is the implementation of sanitation programs after each production cycle [4]. The risk of existing contamination requires not only thorough implementation of sanitation program procedures but also tools for routine assessment of environmental hygiene. Monitoring can reveal shortcomings in cleaning and disinfection processes. It is also important to assess the level of contamination before starting the sanitation procedure so that an appropriate sanitation program can be set and critical contamination points identified [5,6].

For effective hygiene monitoring in animal housing, it is therefore desirable to have practical, rapid, and reliable methods that can be readily applied by farmers [5,7,8]. Methods used for environmental hygiene monitoring can generally be divided into non-microbiological and microbiological approaches. The simplest and most commonly applied method is visual assessment [1,3]. Methodology is typically based on a scoring system [1]. However, this approach is inherently subjective [9]. Additionally, the absence of visible macroscopic contamination does not necessarily exclude substantial organic contamination. Consequently, the use of more precise methods is required to obtain reliable and objective results. The gold standard for hygiene monitoring remains microbiological detection using culture-based (plate) methods. However, these approaches are time-consuming and provide delayed availability of results. Therefore, they are less suitable for routine environmental hygiene control [8,10].

Rapid detection methods currently offer significant potential for monitoring environmental cleanliness in animal housing, as they provide immediate results directly in the field following sample collection [5,11]. Among the available approaches are ATP luminometry and mobile flow cytometry. ATP luminometry is based on the bioluminescent measurement of adenosine triphosphate (ATP). The method requires the substrate luciferin and the enzyme luciferase, which catalyzes the bioluminescent reaction between luciferin and ATP in the presence of oxygen, resulting in light emission. The intensity of the emitted light is expressed in relative light units (RLU) and is directly proportional to the amount of ATP released from cells [12]. ATP is present in all living cells, including animal, plant, and microbial cells. Consequently, luminometry primarily reflects the level of organic contamination, which may be of microbial origin to varying extents depending on the characteristics of the tested matrix [2,13]. ATP luminometers detect ATP derived from any type of organic matter, which may be infectious or may serve as a substrate for the growth of pathogenic microorganisms [13].

The practical application of luminometers has been described across various sectors [14,15]. In agriculture, ATP luminometry has been applied in poultry slaughterhouses [16,17], milking systems [18], livestock transport vehicles [19], and animal housing environments, including laboratory facilities [8], broiler farms [10], and calf housing systems [7,20], where it has been proposed as a practical tool for rapid hygiene assessment under field conditions. The application of luminometry has also been investigated by Yi et al. in pig farrowing pens [2]. Johnson et al. studied the effect of selected cleaning steps in the sanitation program on the level of organic contamination using a luminometer [3].

Flow cytometry is primarily used for the rapid and accurate quantification of microbial cells [21]. The method is based on counting individual cells passing through a flow cell, where they are exposed to a laser beam, resulting in light scattering, absorption, and fluorescence. A major advantage of flow cytometry is the ability to distinguish viable cells from non-viable cells and debris, whereas its limitations include variable sensitivity depending on the probes used. Gunasekera et al. describe a high correlation (r ≥ 0.98) between flow cytometry and traditional cultivation methods in milk analysis [22]. The use of flow cytometry in the food industry is further summarized by Zand et al. [23].

For field applications, mobile flow cytometers are available that rely on impedance measurements and alternating current at different frequencies instead of optical and laser systems. This principle allows for the discrimination and enumeration of intact viable cells as well as residues of organic and inorganic material in samples [24]. Intact cells are all cells with an intact cell wall, regardless of the required growth conditions (aerobic or anaerobic, pH, nutrition, salt concentration, temperature, lag time, incubation time, viable but uncultivable, etc.). The application of an impedance-based flow cytometer to assess the effectiveness of disinfection is mentioned by Pîndaru et al., who in their study confirmed an agreement in 96.42% of results with traditional cultivation methods [25]. Unlike luminometry, flow cytometric measurements are not affected by certain components of disinfectants, salts, or temperature, which may increase the applicability of this method for environmental hygiene monitoring [26]. However, limitations include lower sensitivity compared to laboratory-based cytometers and the inability to identify specific microbial groups. The available mobile flow cytometer CytoQuant (Romer Labs Division Holding GmbH) is primarily designed for use in food-processing and slaughterhouse environments [27]. Studies conducted by Romer Labs have demonstrated its practical applicability in the food industry and its validation against traditional culture-based methods [28]. Preliminary data showed that the results of viable cells gained by mobile flow cytometry were very close to the results of plate count methods. Standard deviations were lower (6–9%) in comparison to the plate count technique (10–21%). To date, however, no studies have evaluated the use of this device in animal housing systems and also no studies are devoted to monitoring the particles which the mobile flow cytometer is able to detect beside the viable cells.

Available studies focus mainly on the application of ATP luminometry in evaluating individual sanitation steps or comparing luminometric methods with traditional microbiological approaches. However, limited information is available regarding the comparability of ATP-based luminometry and mobile flow cytometry for environmental hygiene assessment. Due to their distinct detection principles, both methods may reflect different dimensions of environmental contamination. Their application may also contribute to the implementation of optimal environmental hygiene monitoring strategies not only on pig farms.

The aim of this pilot study was to evaluate and compare ATP luminometry and mobile flow cytometry based on different detection principles under practical field conditions in pig farrowing units. To provide contextual validation, the results obtained using rapid detection methods were compared with conventional microbiological cultivation, which is commonly regarded as a reference approach for hygiene evaluation.

Rapid detection methods were applied prior to sanitation procedures in order to characterize the extent and distribution of organic surface contamination before cleaning and disinfection. This approach enabled rapid contamination assessment directly under field conditions without the delay associated with microbiological cultivation.

2. Materials and Methods

Environmental hygiene monitoring was conducted in farrowing units on two selected pig farms (Farm A and Farm B) after the removal of sows and piglets at weaning. The study was performed in 2024. Farm selection was based on comparable farrowing-unit management, similar housing technology, and willingness to allow repeated field sampling under commercial operating conditions. Both farms were conventional production systems with closed herd turnover. Farrowing units consisted of individual sow housing in crates with slatted floors, while piglets were provided with designated heated solid floor areas. Sows were housed in farrowing crates from two days before parturition until weaning at 28 days postpartum, after which they were transferred to insemination units.

This study was designed as an exploratory proof-of-concept field investigation. Two commercial pig farms were intentionally selected to enable direct comparison of rapid hygiene monitoring methods under practical operating conditions. The limited number of farms reflected the pilot design and the logistical demands associated with repeated on-site measurements using multiple rapid detection devices. The study was designed to evaluate the feasibility and comparative performance of selected rapid detection methods under real farm conditions rather than to provide statistically representative conclusions for pig production systems in general.

2.1. Sampling Conditions and Measurement Procedure

Sampling was conducted on two farms at three defined locations (Figure 1): (1) the heated piglet floor, (2) the farrowing pen wall (sampled 10 cm above the floor), and (3) the floor of the service corridor in the farrowing unit. For monitoring purposes, six farrowing pens were selected in each farrowing unit (marked in red). The capacity of the farrowing unit was identical on both farms (12 farrowing pens per unit). Measurements were performed by a single operator following the manufacturers’ standardized protocols [28,29,30].

Measurements were performed under routine farm operating conditions; therefore, environmental variables such as temperature, humidity, and surface characteristics could not be fully standardized. To reduce variability, identical sampling procedures, sampling areas, and pre-sanitation conditions were applied across both farms. No extreme environmental fluctuations were recorded during the sampling period according to routine farm environmental monitoring systems.

Surface material characteristics and functional use of sampled locations were considered during interpretation of the results, as these factors may influence organic residue accumulation, ATP recovery, and particle detection efficiency.

2.2. Methods and Luminometry Conditions

Three different luminometric devices were used for measurement (Figure 2):

LUM 1: Clean-Trace™ LM1 luminometer (3M Health Care, St. Paul, MN, USA).

LUM 2: EnSURE luminometer (Hygiena LLC, Camarillo, CA, USA).

LUM 3: SystemSURE Plus luminometer (Hygiena LLC, Camarillo, CA, USA).

No cross-device calibration procedure was performed, as substantial inter-device variability among luminometer systems has been widely described in previous studies [31,32]. To improve methodological comparability, identical sampling areas, standardized swabbing procedures, comparable sample handling, and consistent measurement timing were used throughout the study. Therefore, the study focused primarily on comparative trends and contamination differences between sampling locations and devices rather than direct equivalence of absolute RLU values across luminometer systems.

Using luminometers (hereafter LUM), total ATP was determined. For LUM 1, Clean-Trace Surface ATP Test Swabs UXL100 were used (3M Health Care, St. Paul, MN, USA). For LUM 2 and LUM 3, Ultrasnap swab devices were applied (Hygiena LLC, Camarillo, CA, USA). Samples were collected from a defined surface area (100 cm²). During sampling, the swab was rotated to ensure maximum contact with the sampled surface. After collection, the samples were activated in a reagent containing luciferase (according to the manufacturer’s protocol) and measured using the respective device. The results were expressed in RLU per 100 cm².

2.3. Methods and Conditions of Mobile Flow Cytometry

A mobile flow cytometer (MFC)—CytoQuant (Romer Labs Division Holding GmbH, Tulln, Austria)—was used for measurement. In addition to the determination of viable cells (intact cells), the device also enables the quantification of residues (other particles). Detected residual particles may include organic and inorganic material, cellular debris, feed residues, and other surface-associated particulate contamination. Measurements were performed according to the manufacturer’s instructions.

Environmental swab samples were collected from a defined surface area of 100 cm² using a swab and placed into a tube containing a conductive solution (Swab Kit, Romer Labs Division Holding GmbH, Tulln, Austria). After mixing, the tubes were inserted into the instrument, where the sample was aspirated into the flow cytometry chamber. In the course of measurement, cells and particles traversing the detection system were counted using impedance measurement principles. Differentiation is based on conductivity, where viable bacteria exhibit non-conductive properties at low current, in contrast to other residual particles. While the device is capable of quantifying viable intact cells, the present study focused on residual particles in order to assess overall surface contamination prior to sanitation procedures.

This approach allowed more direct comparison with ATP luminometry, which similarly reflects ATP-associated organic contamination rather than viable microorganisms alone. Residual particle values were expressed as the number of residues per 100 cm².

2.4. Methods and Conditions of Microbiological Cultivation

Conventional microbiological cultivation was performed using non-selective blood agar (BA) for determination of colony-forming units (CFU). Prior to sampling, swabs were pre-moistened in transport medium and placed into tubes containing 10 mL of the same transport medium (enzymatic casein hydrolysate, NaCl, glucose, and yeast autolysate) after sample collection. Sampling was performed in paired replicates. Samples were transported to the laboratory in a cooled transport box and inoculated within 24 h after collection. Prior to inoculation, the swab contents were homogenized in the transport medium. Subsequently, 0.5 mL of the obtained suspension was inoculated onto Petri dishes using the spread plate method. Individual paired samples were serially diluted to obtain inoculated concentrations within the range of approximately 10⁰–10³ CFU. Inoculated plates were incubated aerobically at 36 °C for 24 h. After incubation, colonies were enumerated using a manual colony counter, and CFU values per 100 cm² were subsequently calculated for individual samples.

Microbiological cultivation was included primarily as a contextual reference method for comparison with rapid detection techniques rather than as the primary endpoint of the study.

2.5. Statistical Analysis

Results obtained from luminometric measurements (LUM 1–3), mobile flow cytometry (MFC), and microbiological cultivation (TMC 36 °C) were statistically analysed using Unistat for Excel 6.5 (Unistat Ltd., London, UK). Basic descriptive statistics (mean, median, and standard deviation) were calculated.

The complete dataset used for the statistical analyses is provided in Supplementary File S1.

Data normality was assessed using the Shapiro–Wilk and Kolmogorov–Smirnov tests, which indicated both normal and non-normal distributions. For standardisation and graphical presentation, luminometric (RLU), mobile flow cytometry (residual particles) and microbiological cultivation (CFU) data were log-transformed. Prior to transformation, a value of 1 was added to all measurements to avoid zero values.

Relationships between rapid detection methods (LUM 1–3 and MFC) and microbiological cultivation (TMC 36 °C) were evaluated using Spearman’s correlation coefficient (rs). Correlation strength was interpreted according to Hinkle et al. [33] as follows: negligible (0–<0.3), low (0.3–<0.5), moderate (0.5–<0.7), high (0.7–<0.9), and very high (0.9–1).

Confidence intervals (95% CI) for Spearman correlation coefficients were estimated using non-parametric bootstrap resampling (5000 iterations). For near-perfect correlations, confidence intervals were approximated using Fisher’s z-transformation.

Differences in contamination levels between LUM 1–3 and between rapid detection methods (LUM and MFC), and among sampling locations, were analysed using one-way analysis of variance (ANOVA) for normally distributed data. When statistically significant differences were detected, Tukey’s HSD post hoc test was applied. For non-normally distributed data, the Kruskal–Wallis test was used, followed by Dunn’s post hoc test. To complement p-value interpretation, effect size was evaluated using eta squared (η²) for ANOVA and epsilon squared (ε²) for Kruskal–Wallis analysis. Effect sizes were interpreted as small, medium, or large according to commonly accepted thresholds (η²: 0.01, 0.06, 0.14; ε²: 0.01, 0.08, 0.26).

Pairwise differences between selected luminometers were additionally evaluated using Cohen’s d effect size metric to assess practical significance. Cohen’s d values were interpreted as small (0.2), medium (0.5), large (0.8), and very large (>1.2) effects. Confidence intervals (95% CI) were calculated for the main pairwise comparisons to assess the reliability of observed differences.

Differences in organic contamination between farms (A and B) were evaluated using Welch’s t-test or the Mann–Whitney U test, depending on data distribution.

A significance level of p < 0.05 was considered statistically significant, while p < 0.01 indicated highly significant differences.

3. Results

In this pilot study, the results obtained using three different luminometers (LUM 1–3), a mobile flow cytometer (MFC), and microbiological cultivation (TMC 36 °C) were compared under operational farm conditions. Relationships between rapid hygiene monitoring methods and microbiological cultivation were also assessed.

Monitoring was conducted in farrowing units on two pig farms (A–B) after animal removal and prior to sanitation procedures. Samples were collected from defined locations on both farms (corridors, heated pads, and pen walls). In total, 144 samples were collected using rapid detection devices with paired samples additionally analysed by microbiological cultivation.

3.1. Rapid Hygiene Monitoring Methods and Their Comparison

The level of organic contamination was monitored using LUM 1—Clean-Trace™ LM1 luminometer (3M Health Care, St. Paul, MN, USA), LUM 2—EnSURE luminometer (Hygiena LLC, Camarillo, CA, USA), LUM 3—SystemSURE Plus luminometer (Hygiena LLC, Camarillo, CA, USA), and MFC—mobile flow cytometer CytoQuant (Romer Labs Division Holding GmbH, Tulln, Austria).

In this pilot study, the potential application of luminometers (LUM) and a mobile flow cytometer (MFC) for environmental hygiene assessment was evaluated. The relationship between measurements obtained by individual rapid detection devices was analysed using Spearman’s correlation coefficient (rs), separately for each farm and each sampling location (corridors, heated pads, and pen walls) prior to the sanitation procedure. A comparison of results from LUM 1–3 at individual sampling sites on farms A and B is presented in Figure 3.

On Farm A, positive correlations were observed among all luminometers (LUM) across all monitored sampling locations. Statistically significant correlations were found only between LUM 2 and LUM 3 (p < 0.05). A very high correlation was confirmed in the corridors (rs = 1.0000; p < 0.01; 95% CI [0.84, 1.00]), indicating a highly consistent response pattern between both devices. Similarly, very high correlations were observed on heated pads (rs = 0.9429; p = 0.0048; 95% CI [0.56; 0.99]). A high correlation was also observed on pen walls (rs = 0.8857; p = 0.0188; 95% CI [0.26; 0.99]), supporting a high level of agreement between both devices across different surface types. Correlations between LUM 1 and LUM 2, as well as between LUM 1 and LUM 3, were not statistically significant (p > 0.05).

A similar trend was observed on Farm B. Statistically significant correlations were confirmed between LUM 2 and LUM 3 at all sampling locations. A very high correlation was found on pen walls (rs = 0.9429; p = 0.0048; 95% CI [0.56; 0.99]). High correlations were observed on heated pads (rs = 0.8286; p = 0.0416; 95% CI [0.05; 0.98]) and in corridors (rs = 0.8117; p = 0.0499; 95% CI [0; 0.98]). Correlations between LUM 1 and LUM 2, and between LUM 1 and LUM 3 were negative in the case of heated pads and pen walls; however, they were not statistically significant (p > 0.05).

Based on the correlation results, LUM 2 and LUM 3 demonstrated a consistently high level of agreement across different surface types on both farms. In contrast, LUM 1 showed greater variability in relation to the other devices, reflected in higher absolute measurement differences. These findings suggest that luminometers may differ substantially in sensitivity and response characteristics despite measuring the same type of organic contamination.

Differences in measured log RLU values between devices (LUM 1–3) were further evaluated across farms and sampling locations. The mean values of all measurements at the selected sites (corridors, heated pads, and pen walls) prior to sanitation on both farms are presented in

Table 1. Mean log RLU values measured by luminometers (LUM 1–3) at monitored sampling sites on Farms A and B prior to sanitation.

Farm A
Sampling Sites	LUM 1	LUM 2	LUM 3
	Mean Log Value
Corridor	3.91 a	3.71 ab	3.47 b
Heated pads	3.61 a	3.58 a	3.28 b
Pen walls	3.80 ab	3.59 bc	3.22 c
Farm B
Sampling Sites	LUM 1	LUM 2	LUM 3
	Mean Log Value
Corridor	3.76 a	2.32 ab	1.98 b
Heated pads	4.37 a	3.40 b	3.14 c
Pen walls	4.05 ab	3.58 bc	3.28 c

a–c Mean log value in the same sampling site with different superscripts in used devices differ (p < 0.01).

.

Analysis of variance confirmed significant differences in mean log RLU values between luminometers on Farm A in corridors (F(2,15) = 9.35, p = 0.0023, η² = 0.55), heated pads (F(2,15) = 30.47, p < 0.001, η² = 0.80), and pen walls (F(2,15) = 18.72, p < 0.001, η² = 0.71). Similarly, significant differences between luminometers were confirmed on Farm B in corridors (F(2,15) = 684.52, p < 0.0001, η² = 0.99), heated pads (F(2,15) = 8.60, p = 0.0033, η² = 0.53), and pen walls (F(2,15) = 18.72, p < 0.001, η² = 0.71). These η² values indicate a very large effect of device type on measured contamination values.

On Farm A, the highest log RLU values across all monitored sampling locations (corridors, heated pads, and pen walls) were consistently recorded by LUM 1, whereas the lowest values were detected by LUM 3 (p < 0.01). Extremely large practical differences between LUM 1 and LUM 3 were confirmed in corridors (Cohen’s d = 2.58), heated pads (Cohen’s d = 3.58), and pen walls (Cohen’s d = 2.94), indicating substantial differences in absolute contamination values recorded by both devices. No statistically significant differences were found between LUM 1 and LUM 2 at the monitored locations (p > 0.05). Likewise, no statistically significant differences were observed between LUM 2 and LUM 3 in corridors and pen walls (p > 0.05).

Similarly, on Farm B, LUM 1 recorded the highest values across all monitored sampling locations, whereas LUM 3 consistently showed the lowest values (p < 0.01). Very large to extremely large practical differences between LUM 1 and LUM 3 were observed in corridors (Cohen’s d = 16.90), heated pads (Cohen’s d = 11.97), and pen walls (Cohen’s d = 2.25). In the case of heated pads, measurements obtained by all luminometers differed significantly (p < 0.01). No statistically significant differences were detected between LUM 1 and LUM 2, nor between LUM 2 and LUM 3, in corridors and pen walls.

Overall, the results indicate differences in the absolute values measured by individual luminometers despite the high correlations observed between devices. The observed Cohen’s d values further demonstrate that these differences were not only statistically significant but also practically substantial under field conditions.

Subsequently, relationships between LUM and MFC measurements were evaluated. A comparison of results from LUM 1–3 with MFC across individual sampling locations on Farms A and B is presented in Figure 4.

Correlations between LUM 1–3 and MFC were inconsistent across farms and sampling locations, and no statistically significant relationships were confirmed (p > 0.05).

Spearman’s correlation coefficients (rs) ranged from low negative to low positive associations (rs = −0.49 to 0.49) on Farm A, without any clearly defined trend across sampling sites. Confidence intervals were consistently wide and crossed zero in all comparisons (95% CI ranging from −0.93 to 0.94), indicating substantial uncertainty and limited robustness of the observed associations.

For example, correlations between LUM 2 and MFC in corridors (rs = −0.4857; p= 0.3287; 95% CI [−0.93, 0.54]) and between LUM 3 and MFC on pen walls (rs = 0.3714; p= 0.1107; 95% CI [−0.61, 0.92]) demonstrated considerable variability without evidence of a stable relationship between methods.

A similar pattern was observed on Farm B. Although positive trends were recorded for LUM 2 × MFC on pen walls (rs = 0.7714; p = 0.0724; 95% CI [−0.11, 0.97]) and LUM 3 × MFC (rs = 0.7143; p = 0.1108; 95% CI [−0.23, 0.97]), these relationships were not statistically significant and were characterized by broad confidence intervals that included zero.

The mean log values for residual particles measured by MFC were, for corridors, 5.48 (Farm A) and 6.93 (Farm B); for heated pads, 5.63 (Farm A) and 6.80 (Farm B); and for pen walls, 5.43 (Farm A) and 6.96 (Farm B). Across all sampling locations on both farms, MFC consistently produced higher log-transformed values compared to LUM 1–3.

The differences in mean log values relative to LUM 1–3 were, for corridors, 1.78 (Farm A) and 4.25 (Farm B); for heated pads, 2.14 (Farm A) and 3.17 (Farm B); and for pen walls, 1.89 (Farm A) and 3.32 (Farm B). MFC measurements were significantly higher (p < 0.01) than those obtained by LUM 1–3 across all locations. The only exceptions were comparisons between LUM 1 and MFC, where no statistically significant differences were observed (p > 0.05) for corridors on Farm B (LUM 1: 3.76 vs. MFC: 6.93), heated pads on Farm B (LUM 1: 4.37 vs. MFC: 6.80), and pen walls on both Farm A (LUM 1: 3.70 vs. MFC: 5.42) and Farm B (LUM 1: 4.05 vs. MFC: 6.96).

Across all monitored sampling locations, MFC consistently produced substantially higher values than luminometers, reflecting major methodological differences between ATP-based luminometry and impedance-based flow cytometry. Extremely large practical differences between LUM 1 and MFC were confirmed on both farms, with Cohen’s d values ranging from 8.46 to 44.15. On Farm A, Cohen’s d values reached 25.22 in corridors, 22.90 on heated pads, and 8.46 on pen walls. Similarly, on Farm B, extremely large differences were observed in corridors (Cohen’s d = 44.15), heated pads (Cohen’s d = 10.47), and pen walls (Cohen’s d = 11.93). These exceptionally high effect sizes further support the substantial methodological divergence between ATP-based luminometry and mobile flow cytometry.

Effect size analysis using Cohen’s d was performed only for comparisons between LUM 1 and MFC, as these devices demonstrated the greatest absolute differences in measured values across monitored sampling locations. The inclusion of additional pairwise comparisons involving LUM 2 and LUM 3 was considered methodologically redundant because the observed trends between luminometers and MFC were consistent across all ATP-based devices. Therefore, representative comparisons between LUM 1 and MFC were selected to illustrate the magnitude of methodological differences between ATP-based luminometry and impedance-based flow cytometry without unnecessary overextension of statistical outputs.

Overall, the results demonstrated substantial differences between measurements obtained using ATP-based luminometry and mobile flow cytometry under field conditions.

The absence of statistically robust correlations together with consistently broad confidence intervals suggests that luminometers and the mobile flow cytometer do not provide directly interchangeable measurements of environmental contamination. These methods reflect different methodological dimensions of residual surface contamination under field conditions.

3.2. Organic Surface Contamination Prior to Sanitation Measured by Luminometry

Organic contamination at selected sampling sites on both farms prior to sanitation was subsequently evaluated using ATP-based luminometry.

For subsequent analyses, measurements obtained using LUM 2 were selected as representative for graphical presentation and comparative evaluation. LUM 2 consistently produced intermediate values relative to LUM 1 and LUM 3. LUM 2 demonstrated comparable contamination trends across monitored sampling locations. In most monitored locations, no statistically significant differences were observed between LUM 2 and the remaining luminometers (p > 0.05), except for heated pads (Table 1). In contrast, LUM 1 systematically yielded higher values, whereas LUM 3 tended to produce lower absolute measurements. Because all luminometers exhibited similar spatial contamination patterns despite differences in absolute signal intensity, inclusion of all datasets would have resulted primarily in redundant graphical outputs without additional biological interpretation. Therefore, LUM 2 was considered representative for ATP-based environmental contamination assessment under field conditions.

The results of organic contamination monitoring in the form of ATP on Farm A using LUM 2 are presented in Figure 5. In addition, differences in the level of organic contamination (ATP) between the monitored farms and sampling locations were evaluated (

Table 2. Mean ATP-based organic contamination (log RLU) measured by luminometry (LUM 2) at monitored sampling sites on Farms A and B prior to sanitation.

Luminometry—ATP
Sampling Sites	Farm A	Farm B
	Mean Log Value
Corridor	3.71 a, x	2.32 b, y
Heated pads	3.58 a, x	3.40 b, x
Pen walls	3.59 a, x	3.58 a, x

a,b Mean log value in the same sampling site with different superscripts between farms differ (p < 0.01). x,y Mean log value in the same farm with different superscripts between sampling site differ (p < 0.01).

).

The results showed that the level of organic contamination in terms of ATP on the farrowing unit of Farm A ranged from 3.46 to 3.89 log RLU (corridors), 3.49 to 3.63 log RLU (heated pads), and 3.47 to 3.69 log RLU (pen walls). Differences in the level of organic contamination were assessed among the selected sampling locations (corridors, heated pads, and pen walls).

The variability in ATP contamination across monitored sampling locations was relatively narrow, indicating a homogeneous distribution of organic residues within the farrowing unit prior to sanitation. The highest mean level of organic contamination was detected in the corridors (3.71 log RLU), while the lowest values were observed on pen walls (3.59 log RLU) and heated pads (3.58 log RLU). Although analysis of variance did not confirm statistically significant differences between monitored sampling sites (F(2,15) = 2.36, p = 0.1281), the observed effect size was large (η² = 0.24), suggesting a potentially meaningful spatial trend in ATP contamination distribution within the farrowing environment.

The results of organic contamination monitoring in terms of ATP on Farm B using LUM 2 are presented in Figure 6.

On the farrowing unit of Farm B, the level of organic contamination (ATP) ranged from 2.28 to 2.36 log RLU (corridors), 3.20 to 3.55 log RLU (heated pads), and 3.19 to 3.89 log RLU (pen walls).

Differences in the level of organic contamination among the selected sampling locations (corridors, heated pads, and pen walls) were further evaluated. In contrast to Farm A, analysis of variance confirmed highly significant differences in ATP contamination between monitored sampling sites on Farm B (F(2,15) = 77.86, p < 0.001). The highest mean level of organic contamination was detected on pen walls (3.58 log RLU) and heated pads (3.40 log RLU), while the lowest was observed in corridors (2.32 log RLU). Post hoc comparisons confirmed highly significant differences between corridors and the remaining sampling sites (p < 0.01), while no statistically significant difference was detected between pen walls and heated pads (p > 0.05). Moreover, the observed effect size was extremely large (η² = 0.912), indicating that sampling site location accounted for the majority of variability in ATP contamination on Farm B.

The results of this pilot study also highlighted differences in the level of organic contamination between Farms A and B at the monitored locations. On Farm B, significantly lower contamination levels were detected in corridors compared to Farm A (p < 0.01). Lower contamination was also observed on heated pads on Farm B (p = 0.0389). No differences in organic contamination levels on pen walls were found between Farms A and B (p > 0.05).

Overall, the results demonstrate that ATP-based contamination patterns may differ considerably between farms and surface types under practical field conditions. These findings highlight the importance of targeted sampling site selection when luminometry is applied for routine hygiene monitoring in farrowing environments.

3.3. Organic Surface Contamination Prior to Sanitation Measured by Mobile Flow Cytometry

The level of organic contamination was assessed based on the presence of residual particles detected by the MFC system. The results of residual particle monitoring on Farms A and B are presented in Figure 7 and Figure 8. In addition, differences in the level of residual particles between the monitored farms and sampling locations were evaluated (

Table 3. Mean residual particulate contamination detected by mobile flow cytometry at monitored sampling sites on Farms A and B prior to sanitation.

Mobile Flow Cytometry—Other Particles
Sampling Sites	Farm A	Farm B
	Mean Log Value
Corridor	5.48 b, y	6.93 a, x
Heated pads	5.63 b, x	6.80 a, x
Pen walls	5.43 b, y	6.96 a, x

a,b Mean log value in the same sampling site with different superscripts between farms differ (p < 0.01). x,y Mean log value in the same farm with different superscripts between sampling site differ (p < 0.01).

).

On Farm A, residual particle levels ranged from 5.39 to 5.58 log in corridors, 5.60 to 5.68 log on heated pads, and 5.41 to 5.45 log on pen walls.

Analysis of variance confirmed highly significant differences in residual particle contamination between monitored sampling sites on Farm A (F(2,15) = 23.03, p < 0.001). The highest level of residual particles was detected on heated pads (5.63 log), which was significantly higher compared to corridors (5.48 log) and pen walls (5.43 log). No statistically significant difference in residual particle contamination was observed between corridors and pen walls (p > 0.05). Moreover, the observed large effect size (η² = 0.754), indicates that sampling site location substantially influenced residual particle contamination detected by mobile flow cytometry.

The results of residual particle monitoring on Farm B are presented in Figure 8.

On Farm B, the level of residual particles on the farrowing unit ranged from higher log values, specifically 6.87–6.98 log (corridors), 6.24–7.00 log (heated pads), and 6.92–7.00 log (pen walls). Heated pads exhibited a wider range of measured residual particle values compared to corridors and pen walls.

The highest mean level of residual particles was detected on pen walls (6.96 log), followed by corridors (6.93 log) and heated pads (6.80 log). However, Kruskal–Wallis analysis did not confirm statistically significant differences in residual particle contamination between monitored sampling locations (χ² = 1.57, p = 0.4563). Furthermore, the observed effect size was negligible (ε² ≈ 0), indicating minimal influence of sampling site location on residual particle contamination detected by mobile flow cytometry on Farm B.

Differences in residual particle levels (Table 3) further confirmed that Farm B showed significantly higher contamination compared to Farm A across all monitored sampling locations (p < 0.01).

Unlike luminometry, mobile flow cytometry quantifies a broader spectrum of particulate contamination independently of ATP activity. Consequently, residual particle measurements likely reflect different dimensions of environmental contamination compared to ATP-based organic residue detection.

3.4. Rate of Microbial Contamination and Comparison with Rapid Detection Methods

To compare the results obtained using rapid detection methods, plate cultivation methods were additionally included. Total mesophilic bacteria cultivated at 36 °C (TMC 36 °C) were monitored. The results of microbial contamination monitoring on Farms A and B are presented in Figure 9 and Figure 10.

The highest level of microbial contamination on both farms was detected in corridors (Farm A: 4.90–5.08 log CFU; Farm B: 4.88–5.18 log CFU), with these values being significantly higher (p = 0.0001) compared to pen walls and heated pads on both farms. The lowest level of microbial contamination was observed on heated pads (Farm A: 3.27–3.35 log CFU; Farm B: 3.79–4.17 log CFU).

On Farm A, pen wall contamination was significantly higher than contamination on heated pads (p = 0.001), whereas no significant difference between these locations was confirmed on Farm B (p > 0.05).

Analysis of variance confirmed highly significant differences in microbial contamination between monitored sampling sites on Farm A (F(2,15) = 2470.54, p < 0.001). Moreover, the observed effect size was extremely large (η² = 0.997), indicating that sampling site location accounted for nearly all variability in microbial contamination detected by plate cultivation (TMC 36 °C).

Similarly, analysis of variance confirmed highly significant differences in microbial contamination between monitored sampling sites on Farm B (F(2,15) = 124.46, p < 0.001). The observed effect size was also extremely large (η² = 0.943), indicating that sampling site location accounted for the majority of variability in microbial contamination detected by plate cultivation (TMC 36 °C).

Relationships between values obtained using rapid detection methods (LUM 1–3 and MFC) and plate cultivation methods (TMC 36 °C) were subsequently evaluated. Comparisons of results obtained at individual sampling sites on Farms A and B are presented in Figure 11.

The results confirmed correlations between rapid detection methods and plate cultivation (TMC 36 °C) only at selected sampling sites and for specific detection methods. On Farm A, statistically highly significant and very strong correlations were confirmed in corridors for LUM 2 × TMC 36 °C (r_s = 0.9429, p = 0.0048; 95% CI = [0.559; 0.994]) and LUM 3 × TMC 36 °C (r_s = 0.9429; p = 0.0286; 95% CI = [0.559; 0.994]). Similarly, a statistically highly significant very strong correlation was detected on heated pads for LUM 1 × TMC 36 °C (r_s = 0.9429, p = 0.0048; 95% CI = [0.559; 0.994]). In contrast, no statistically significant correlations were confirmed between MFC and TMC 36 °C on Farm A, with correlation coefficients ranging from weak negative to strong negative associations (r_s = −0.20 to −0.95) accompanied by broad confidence intervals.

Farm B demonstrated an absence of statistically significant correlations between LUM 1–3 and TMC 36 °C across monitored sampling sites (p > 0.05). However, a statistically significant strong positive correlation was confirmed between MFC and TMC 36 °C on heated pads (r_s = 0.8452, p = 0.0341; 95% CI = [0.107; 0.983]). Although moderate positive correlations were additionally observed between MFC and TMC 36 °C in corridors and on pen walls (r_s = 0.6571), these relationships were not statistically significant and were associated with wide confidence intervals.

Overall, correlations between rapid detection methods and plate cultivation varied considerably between farms and sampling locations.

4. Discussion

Despite the use of ATP luminometry in livestock environments, several limitations remain unresolved. Most published studies focus on the application of a single luminometer without direct comparison between different devices under the same field conditions. Furthermore, the application of mobile flow cytometry (MFC) in livestock housing environments remains largely unexplored. Consequently, the relationship between ATP-based and particle-based rapid detection methods, as well as their agreement with microbiological cultivation, is still insufficiently understood.

The present pilot study comparatively evaluated ATP luminometry and mobile flow cytometry for environmental hygiene monitoring in pig farrowing units. The obtained data were additionally compared with microbiological cultivation (TMC 36 °C) to provide contextual validation. The study contributes to the understanding of the applicability, limitations, and interpretative value of rapid hygiene monitoring methods in commercial pig production systems.

Specifically, the study aimed to:

• Assess the agreement between different ATP luminometers used for environmental hygiene monitoring;
• Compare ATP-based luminometry with mobile flow cytometry in the assessment of surface contamination;
• Evaluate relationships between rapid detection methods and microbiological cultivation (TMC 36 °C);
• Evaluate whether these rapid detection methods provide overlapping or complementary information on environmental contamination;
• Determine the applicability of these methods for identifying potentially critical contamination sites prior to sanitation procedures in commercial pig farms.

4.1. Potential of Rapid Hygiene Monitoring Methods for On-Farm Use and Their Relationship to Microbiological Cultivation

Selection of an appropriate hygiene monitoring method in livestock production should consider the specific environment conditions under which monitoring is carried out. Livestock housing systems differ in housing technology, cleaning and disinfection procedures, surface materials, and contamination burden. These factors may influence the ability of devices to detect contamination, as reported by Turner et al. [8].

Environmental contamination in livestock housing systems consists of heterogeneous mixtures of organic and inorganic material, which may influence the interpretation of rapid hygiene monitoring results [34,35,36].

In this study, ATP luminometry was used to assess ATP-associated organic contamination, whereas mobile flow cytometry (MFC) was applied for detection of residual particulate contamination [28,29,30]. Both methods demonstrated applicability for monitoring environmental contamination under livestock housing conditions. Figure 12 provides a schematic representation of the spectrum of contaminants detected by these methods.

Luminometry (LUM), based on ATP measurement, provides information on organic contamination regardless of its origin. ATP measurements do not directly reflect the level of microbial contamination [36]. RLU values may represent the presence of intact microorganisms, but also dead cells and other organic material present in the environment [3,18,31].

Previous studies have reported positive correlations between ATP luminometry and microbiological cultivation under conditions of high contamination burden [7,20]. Some studies report that high CFU counts correlate with high RLU values, and the presence of organic contamination is associated with a higher risk of microorganisms on surfaces [32,37].

Nevertheless, LUM measurements should primarily be interpreted as indicators of organic contamination rather than purely microbial contamination, as also stated by Bakke and Suzuki [38]. The observed correlations are strongly influenced by the level of surface contamination present on monitored surfaces. Therefore, correlations between ATP measurements and microbial counts should not be considered a basis for replacing traditional culture methods with LUM, as also noted by Johnson et al. [3]. Studies such as Ching et al. [39] have shown that ATP values do not necessarily correspond to the number of viable microorganisms present on surfaces. Disinfectants break down the cell walls of animals, plants and microorganisms. It means that those cells die while their ATP is preserved. The result is high counts that do not reflect the actual degree of bacterial contamination.

MFC differs from ATP luminometry in its ability to detect viable cells (intact cells) as well as residual particulate material beyond ATP-containing organic contamination [27]. In this pilot study, residue detection (“other particles”) was evaluated. These particles are not identified in the detection chamber as intact living bacterial cells, but represent broader particulate contamination present on monitored surfaces.

The results demonstrated substantial methodological differences between ATP-based and particle-based contamination assessment. Across all monitored sampling locations, MFC consistently produced markedly higher contamination values compared to luminometry. These differences are likely related to the distinct contamination spectrum detected by each method. ATP luminometry primarily reflects ATP-associated organic contamination, whereas MFC quantifies a broader range of particulate material independently of ATP activity. These findings were further supported by large effect sizes between methods. Cohen’s d values ranged from 8.46 to 44.15.

The absence of consistent correlations between luminometry and MFC further supports the assumption that both methods characterize different aspects of environmental contamination. From the schematic representation, it is evident that residues detected by MFC include, in addition to viable cells, all organic material containing ATP, thereby partially overlapping with the detection scope of LUM. The amount and composition of organic material present on monitored surfaces may therefore influence the degree of agreement between both methods. Unlike luminometry, however, MFC is also capable of detecting inorganic particles.

The observed lack of agreement between rapid methods was further supported by comparison with microbiological cultivation (TMC 36 °C). Detected variability among farms and sampling locations reflects the heterogeneous composition of environmental contamination within livestock housing systems. Corridors showed the highest microbial contamination on both farms, which may be associated with intensive animal movement and transfer of organic material within the housing environment. In contrast, elevated residual particulate contamination detected by MFC was observed particularly on heated pads, where feed residues, biological material, and accumulated surface deposits may contribute to increased particulate contamination.

Despite their methodological differences, both ATP luminometry and MFC can be considered useful tools for rapid screening of environmental contamination in livestock production systems. Although both methods were originally developed for food industry applications, the present study demonstrated their applicability under livestock housing conditions characterized by higher contamination loads. It should be noted that in higher-risk environments such as slaughterhouses, ATP-based contamination measured by LUM may be even higher [16]. However, surface hygiene is often not assessed using these rapid methods prior to sanitation in such settings.

ATP luminometry enables rapid assessment of ATP-associated organic contamination and overall surface cleanliness [1,19,40]. In contrast, MFC may provide complementary information regarding broader particulate contamination, including contamination not directly associated with ATP-containing biological material [27,28].

This pilot study demonstrated that LUM and MFC represent applicable methods for rapid on-site hygiene assessment under pig housing conditions. Both methods enable on-site measurement without the need for laboratory sample processing, which is a key advantage for routine hygiene monitoring. However, correct interpretation requires understanding the measurement principles and influencing factors described above. Rapid detection methods should therefore be considered complementary tools for environmental hygiene assessment rather than direct replacements for conventional microbiological cultivation.

4.2. Type of Luminometer as a Factor Influencing Measurements of Organic Contamination

The results of this study demonstrated that the type of luminometer substantially influenced measured ATP contamination values under field conditions. Differences between luminometers likely reflected variations in device sensitivity, detection chemistry, swab systems, and signal processing. ATP luminometry therefore appears more suitable for relative within-device hygiene monitoring rather than direct comparison of absolute RLU values between different luminometer systems.

A range of devices from different manufacturers is available for total ATP measurement, differing in their specifications and suitability for various conditions [3,12]. Luminometers are used to identify critical contamination points, validate cleaning procedures, monitor cleaning efficiency, and support staff training in hygiene practices [32]. These applications are relevant not only in the food industry but also in animal production systems, as demonstrated in this pilot study. However, results obtained with different devices are generally not directly comparable [31,32]. This is due to differences in detection sensitivity as well as device-specific characteristics [3,15].

An important limitation of ATP luminometry is the absence of standardized RLU units within the SI system. Consequently, RLU values obtained from different luminometers may differ substantially, which should be considered when comparing results across studies and during routine farm monitoring [32].

In this pilot study, three commonly used luminometers were compared under practical livestock housing conditions [3,19]. These included LUM 1—Clean-Trace™ LM1 (3M Health Care, USA) with a detection limit of 3 fmol ATP [32], LUM 2—EnSURE (Hygiena LLC, USA) with a detection limit of 1 fmol ATP, and LUM 3—SystemSURE Plus (Hygiena LLC, USA) also with a detection limit of 1 fmol ATP. The devices differ not only in detection limits but also in measurement range. However, a wider measurement range does not necessarily indicate higher sensitivity, as also highlighted by studies evaluating luminometer performance [41]. LUM 3 showed the largest range of measured values (2.0 log), while LUM 1 exhibited the most consistent results (1.09 log) across all sampling sites and farms.

Measurement variability is also influenced by the sampling system used. Swab systems differ in design, which may affect the sampling procedure and efficiency of material collection. Standardization of swabbing is therefore essential. Gamo [42] highlights that differences in applied pressure during swabbing can significantly affect results. The use of templates is also recommended, and adherence to manufacturer protocols is crucial [29,30].

The observed agreement between luminometers was greatest for devices sharing similar technical characteristics and sampling systems. These findings support previous reports indicating that device construction, reagents, and sampling systems substantially influence ATP measurements [31,32]. Consequently, ATP values obtained using different luminometer types should not be interpreted using universal reference thresholds without considering device-specific characteristics.

From a practical perspective, luminometric data should be interpreted primarily in a relative manner—within the same device, farm, or monitoring period—rather than based on absolute values. Manufacturers also highlight the importance of establishing site-specific critical thresholds rather than relying on generalized criteria [26]. For routine hygiene monitoring, the ability of LUM to detect changes in organic contamination over time or between locations is therefore more important than the absolute RLU value itself.

Based on the observed differences and correlations between devices, LUM 2 was selected for detailed evaluation of spatial contamination patterns because it represented an intermediate measurement range and showed higher consistency with the third device. This approach minimized instrument-related variability and allowed a more reliable comparison of sampling locations and farms.

Overall, the results indicate that devices with similar detection sensitivity show greater agreement across conditions, while differences become more pronounced between instruments with differing sensitivity. Nevertheless, differences between devices do not preclude their practical application for farm hygiene monitoring, provided that results are interpreted within the context of the specific device used. Proper analysis of data must always follow manufacturer protocols. In livestock environments, it is advisable to establish site-specific threshold scales for each luminometer type, similarly to practices in the food industry [1]. Such scales can be developed through repeated sampling over time in a given facility, as described by Lindell et al. [18]. High measured values should serve as an indicator for reassessment of sampling sites, repeat sampling, and consideration of microbiological culture analysis [14].

The comparative evaluation of luminometers with differing technical characteristics under identical field conditions therefore represents an important methodological contribution to interpretation of ATP-based hygiene monitoring in livestock production systems.

4.3. Organic Contamination Level on Farms and Identification of Critical Points Prior to Sanitation

Assessment of the initial level of organic contamination before sanitation may provide information on the overall hygiene status of the farrowing environment and help identify critical points that require sufficient attention during cleaning and disinfection procedures [43]. Previous studies identified feeders, drinkers, and entry floors as critical contamination points due to insufficient cleaning efficiency and accumulation of organic material [1,2,3]. In contrast, farrowing pen walls are considered more easily cleanable surfaces [44]. Surface characteristics may influence both inorganic and organic contamination and thereby affect microbial persistence and survival [5].

In this study, on both farms sampling sites (corridors, heated pads, and pen walls) were selected based on differences in material, surface characteristics, and level of animal contact. The aim was to identify potentially high-risk areas for infection transmission as well as locations that are more difficult to clean. The findings indicate that contamination prior to sanitation is not homogeneously distributed within farrowing units and may vary considerably according to surface characteristics and farm-specific operating conditions. Rapid hygiene monitoring methods enabled identification of locations with elevated contamination levels that may represent potential hotspots for persistence of organic material and subsequent microbial survival.

Significant differences between farms were particularly evident in corridors, which are surfaces frequently contaminated by personnel movement within the farrowing units. Despite sanitation procedures, these areas may remain highly contaminated. These surfaces are therefore not the best indicators of overall farrowing unit cleanliness, as also noted by Johnson et al. [3], but they represent critical points for potential recontamination of pens. Initial contamination levels should therefore be considered during sanitation procedures, especially due to surface porosity, which facilitates accumulation of contamination.

Heated pads may represent the most contaminated surface prior to sanitation due to contamination with urine, faeces, and feed residues. In contact with piglets, this surface can be considered microbiologically high-risk, as it may provide optimal conditions for microbial growth [45]. Therefore, heated pads require targeted intervention during sanitation, including enhanced mechanical cleaning and rinsing to prevent adherence of organic material.

Pen walls showed contamination levels comparable to other locations before sanitation (Farm A similar to corridors; Farm B similar to both corridors and heated pads). Their smooth surface reduces the risk of contamination persistence after cleaning.

Differences between farms were also observed, suggesting that farm-specific management and environmental conditions may influence contamination burden prior to sanitation.

MFC results complement luminometric assessments and demonstrate that contamination distribution may also differ in terms of total surface burden of residual particles, which additionally include inorganic components. The different detection principles of both methods resulted in distinct value distributions and the absence of a consistent correlation between LUM and MFC. These differences confirm that the methods capture different fractions of contamination and provide complementary information on hygiene status.

Cultivation-based microbial assessment (TMC 36 °C) demonstrated only partial agreement with rapid detection methods. The highest microbial contamination levels on both farms were detected in corridors, whereas the lowest microbial contamination levels were observed on heated pads. These findings contrasted particularly with MFC measurements on Farm A, where heated pads represented the most contaminated surfaces in terms of residual contamination.

These findings indicate that LUM, MFC, and cultivation-based methods assess different aspects of environmental contamination and should not be interpreted as interchangeable approaches. ATP luminometry primarily reflects ATP-associated organic residues, MFC detects residual particulate contamination including inorganic material, and cultivation methods assess viable microbial contamination.

Overall, the results indicate that identification of critical contamination sites prior to sanitation cannot rely on generalized assumptions alone but should consider farm-specific conditions, including housing technology, operational management, animal activity, and surface characteristics. Rapid detection methods may therefore represent useful tools for identification of contamination hotspots prior to sanitation, particularly when interpreted in combination with cultivation-based microbiological assessment.

4.4. Study Limitations, Contribution, and Future Research

This study has several limitations that should be considered when interpreting the results. The study was designed as a pilot exploratory field investigation and included only two commercial pig farms. Although this design enabled repeated application and direct comparison of rapid hygiene monitoring methods under practical operating conditions, the limited number of farms restricts the generalizability of the findings to broader pig production systems. Farm-level effects such as differences in personnel movement, surface wear, cleaning routines, microclimate, and animal-related contamination dynamics may have influenced the observed contamination patterns.

Measurements were performed under routine commercial farm conditions, where environmental variables such as temperature, humidity, surface porosity, and material characteristics could not be fully standardized. Although identical sampling procedures, sampling areas, and pre-sanitation conditions were applied across farms, uncontrolled environmental variability may have influenced ATP recovery, particulate detection, and microbiological cultivation results.

Another limitation relates to the comparison of rapid detection methods based on different detection principles. ATP luminometry primarily reflects ATP-associated organic contamination, whereas mobile flow cytometry additionally detects a broad spectrum of particulate residues independent of microbial viability. Consequently, direct comparison between methods should be interpreted cautiously, as the evaluated techniques likely characterize different dimensions of environmental contamination rather than identical contamination fractions.

The study focused primarily on residual particulate contamination detected by the mobile flow cytometer rather than viable cell quantification. Although this approach enabled more direct comparison with ATP-associated organic contamination, inclusion of viable cell analysis could provide additional information regarding relationships between particulate contamination and microbiological viability under field conditions.

Another methodological limitation is the absence of cross-device calibration between luminometers. However, substantial inter-device variability among luminometric systems has been widely described in previous studies. For this reason, the present study focused primarily on comparative trends and relative contamination differences between devices and sampling locations rather than direct equivalence of absolute RLU values.

The relatively small sample size also contributed to broad confidence intervals in several correlation analyses, indicating limited robustness of some observed associations. Consequently, the results should be interpreted primarily as exploratory observations intended to evaluate methodological applicability and comparative behaviour of rapid hygiene monitoring methods under commercial farm conditions.

Despite these limitations, the study provides additional insight into the practical applicability and interpretative limitations of rapid hygiene monitoring methods under livestock housing conditions. The findings further identified potential critical sites (corridors), where a higher persistence of contamination prior to sanitation may occur. This may be valuable for optimizing sanitation protocols and improving hygiene management in specific farm conditions.

Future studies should include a larger number of farms, repeated longitudinal sampling, seasonal variability assessment, and expanded microbiological validation, including viable cells and pathogen detection. Further standardization and calibration approaches may additionally improve comparability between rapid hygiene monitoring systems and support the establishment of practical interpretation thresholds for field application. Future research should also focus on comparing rapid detection methods with standard plate culture techniques to better understand the relationship between organic contamination, residual particles, and the presence of microorganisms. Another area of interest is the evaluation of rapid methods for monitoring the effectiveness of individual cleaning and disinfection procedures.

Overall, this pilot study provides proof-of-concept evidence that ATP luminometry and mobile flow cytometry can be implemented under practical livestock farm conditions. The findings support their use as complementary rapid hygiene monitoring tools requiring context-specific interpretation rather than universal contamination thresholds.

5. Conclusions

This pilot study demonstrated that ATP-based luminometry and mobile flow cytometry can be practically implemented for rapid hygiene monitoring in pig farrowing units under commercial farm conditions. Both methods provide immediate information on surface contamination levels prior to sanitation procedures without the need for laboratory processing of samples, which is a key advantage for routine hygiene control in livestock production.

The study confirmed that contamination prior to sanitation is not homogeneously distributed within farrowing units and varies according to sampling location and farm-specific conditions. Corridors represented important sites of elevated microbiological contamination detected by cultivation methods, whereas mobile flow cytometry identified increased residual particulate contamination particularly on heated pads and pen walls. Rapid detection methods enabled identification of potential contamination hotspots requiring increased attention during cleaning and disinfection procedures.

Substantial differences were observed between luminometers despite the shared ATP-based detection principle. These findings emphasize that ATP measurements obtained using different luminometers should not be interpreted as directly interchangeable without careful consideration of device-specific characteristics.

The study further demonstrated that ATP luminometry, mobile flow cytometry, and cultivation-based microbial assessment characterize different aspects of environmental contamination and should therefore be interpreted as complementary rather than interchangeable approaches.

From a practical perspective, the findings indicate that rapid detection methods may represent useful complementary tools for routine hygiene monitoring and identification of potentially critical contamination sites prior to sanitation. However, interpretation of results should remain method-specific and farm-specific due to substantial environmental variability and methodological differences between detection principles.

Luminometry and mobile flow cytometry may contribute to improved biosecurity, reduced risk of infectious disease transmission, and enhanced animal welfare. As previously noted, each production system has specific management and housing conditions; therefore, contamination thresholds established in one farm cannot be directly applied to others. For this reason, it is recommended that a sufficient number of measurements be performed prior to evaluation in each farm in order to establish a site-specific cleanliness scale.

The results of this study support their inclusion as part of a holistic hygiene monitoring approach combining visual inspection, rapid screening methods, and targeted microbiological assessment.