Analyzing Salmonella, Campylobacter, and generic Escherichia coli in feed ingredients and poultry feed samples obtained from commercial and integrated feed mills

Abstract

Microbial contamination in feed mills poses a significant threat to biosecurity and food microbial safety. Understanding pathogen occurrence in feed mills is crucial for minimizing contamination and producing hygienic feed. The objective of this study was to determine the prevalence of Salmonella, Campylobacter, and generic Escherichia coli (E. coli) in raw ingredients and finished feed from feed mills across the southeastern United States. A total of 180 samples were obtained from 14 commercial and integrated feed mills in February 2024 and February 2025. Samples were analyzed using culture methods, with presumptive-positive Salmonella and Campylobacter isolates confirmed by conventional Polymerase Chain Reaction (PCR). Compact Dry Plates were used to evaluate E. coli (EC), Total Counts Rapid (TCR), and Coliforms (CF). Data were analyzed using PROC GLIMMIX to evaluate the effects of feed mill facility, sample type, and year on pathogen prevalence. Results indicated that EC, TCR, and CF were prevalent across all mills and sample types. Twenty-three samples were presumptive-positive for Salmonella, and all were confirmed by PCR. Additionally, 42 samples were presumptive-positive for Campylobacter (23.33%); however, only 14 were confirmed by PCR (7.78%). These findings demonstrate that feed manufacturing environments remain potential sources of bacterial contamination and highlight the value of pathogen surveillance beyond routine indicator organisms. The recovery of confirmed Campylobacter further emphasizes the importance of including this pathogen in feed safety monitoring programs.

Introduction

As of 2024, the United States poultry industry has reached a record-high production level, with the United States Department of Agriculture (USDA) reporting approximately 27 billion kilograms of chicken produced, valued at around $45 billion (USDA, 2024). Poultry is the most consumed protein in the United States and remains an important source of foodborne illness. The most recent Interagency Food Safety Analytics Collaboration (IFSAC) report attributes chicken to approximately 19% of foodborne salmonellosis cases in the United States, the highest percentage among specific food categories (IFSAC, 2026). Salmonella and Campylobacter are the two most common foodborne pathogens associated with human foodborne illness from poultry consumption and are a concern to feed manufacturers because contamination of feed ingredients and feed manufacturing environments can introduce and disseminate these pathogens within the poultry production chain (Stanley and Bajagai, 2022).

Poultry integrators in the United States are continually working to minimize cross-contamination throughout all stages of production, ensuring a safe, wholesome, and cost-effective protein source is available to consumers (NCC, 2020; Obe et al., 2020). Feed mills play a critical role in the colonization-infection-contamination cycle, as contaminated feed ingredients and feed manufacturing environments can serve as sources of pathogen introduction into poultry flocks and contribute to contamination throughout the production chain (Jones, 2011; Huss et al., 2015; Cochrane, 2016). In 2011, the implementation of the Food Safety Modernization Act (FSMA) caused a pivotal shift in feed mill operations. Under the FSMA, all stages of the food supply chain for both human and animal consumption must comply with safety standards established by the United States Food and Drug Administration (FDA; FDA, 2024). These standards encompass implementing current good manufacturing practices (cGMPs), developing hazard analyses with preventive control plans (HACCP), and implementing a validation process with accompanying documentation to ensure hazards are effectively controlled (FDA, 2024).

Among these standards, GMPs in feed mills, such as evaluating facility design and layout, maintaining employee hygiene, and implementing an effective pest and rodent management plan, are long-standing practices for controlling microbial contamination (Huss et al., 2015). Recognizing that these practices are important, it is not only crucial to implement strategies to prevent microorganisms from entering the facility, but also to apply control interventions to reduce microbial multiplication and eliminate existing microorganisms through physical, chemical, or mechanical treatments (Steghöfer et al., 2021).

It is well established that the receiving area, where cereal grains and raw ingredients are unloaded, contains some of the highest microbial loads within a feed mill. This is largely attributed to microorganisms associated with incoming ingredients and to the movement of dust, vehicles, and personnel through the receiving area, making it a primary route for introducing microorganisms into the feed mill environment (Davies and Wray, 1997; Jones, 2011). Biosecurity plans and measures are implemented in feed mills to reduce foot traffic and microbial cross-contamination (Huss et al., 2018). However, even in areas where biosecurity is enforced, cross-contamination within feed mill equipment can occur, potentially contaminating subsequent batches of feed (Davies and Wray, 1997; Cochrane, 2016; Gosling et al., 2022). Globally, there is an abundance of research evaluating pathogen presence in feed manufacturing facilities; however, most of these studies focus on the finished feed product, and there is a gap in the literature evaluating pathogen detection within the feed mill environment (Magossi et al., 2019; Parker et al., 2022; Gosling et al., 2022). Screw conveyors, pellet coolers, storage bins, and elevator boot pits are among the highest-risk pieces of equipment for microbial contamination, but due to facility design, they are extremely difficult to clean (Cochrane, 2016). Understanding pathogen occurrence within the feed mill environment is critical, as contamination from multiple stages of feed production can serve as a continual source of exposure for multiple batches of feed.

A comprehensive evaluation of multiple pathogens within feed mills is necessary to capture the complexity of microbial contamination and its implications for poultry health and food safety. The most common microorganisms found in poultry and livestock feeds are Salmonella (Chadwick, 2017; Hacking et al., 1978; Kidd et al., 2002; Maciorowski et al., 2006; Walker, 1959), Escherichia coli (E. coli; Dargatz et al., 2005; Davis et al., 2003), and Clostridium perfringens (Muñoz et al., 2021; Udhayavel et al., 2017). While the severity of each of these varies, microbial multiplication in feed mills depends on moisture levels, feed composition, and thermal processing intensity (Betiku et al., 2025; Jones, 2011). Evaluating multiple microorganisms in feed is vital to fully characterize microbial contamination within poultry feed and its potential impact on bird health and food safety.

Evaluating E. coli, Total Counts Rapid, and coliforms provides feed mills with a comprehensive assessment of hygiene conditions. Coliforms are naturally found in the guts of animals and humans and are typically a source of fecal contamination (Mahami et al., 2019). Although these are typically harmless, evaluating coliform levels in feed mills is beneficial because high levels indicate poor hygiene conditions, which may be linked to fecal contamination.

Escherichia coli is the most well-known coliform, although many strains are naturally occurring. Ngai et al. (2021) evaluated 150 feed samples from various feed types, and 58% were contaminated with E. coli. Total Counts Rapid provides a combination of bacterial, yeast, and mold counts, which are useful for evaluating raw ingredient quality or potential spoilage concerns after the feed has been in storage for an extended period.

Currently, feed mills are not a major focus for Campylobacter control during production stages because the pathogen is sensitive to oxygen and low moisture; therefore, feed is considered an undesirable environment unless previously contaminated (El-Saadony, 2023). Previous research by El Saadony et al. demonstrated that although Campylobacter may not be able to grow and replicate, it can survive in unfavorable conditions (El-Saadony et al., 2023). The absence of routine Campylobacter surveillance in feed mills represents a research gap that must be addressed to protect consumers, as poultry products remain a primary pathway of Campylobacter transmission to humans. Furthermore, Campylobacter's ability to enter a viable but non-culturable (VBNC) state may complicate detection and surveillance efforts within the feed mill.

Therefore, the objective of this study was to test raw ingredients and different feed types from commercial and vertically integrated feed mills in the Southeastern United States for Campylobacter and Salmonella, with positive findings confirmed through PCR analysis. Additionally, the presence of E. coli (EC), total counts rapid (TCR), and coliforms (CF) was evaluated from Compact Dry plates.

Materials and methods

Experimental design

Feed samples were collected from commercial and integrated feed mills across the Southeastern United States. Table 1 summarizes the number of samples collected from each feed mill and the number of samples collected for each feed type in this study. In total, 180 samples were obtained: 106 from February 2024 and 74 from February 2025. The following sample types were collected for this study: raw ingredients (including ground corn, soybean meal, and DDGs), starter, grower, finisher, withdrawal, breeder, and pullet developer feed. All samples were stored in a -20°C walk-in freezer and thawed at room temperature (approximately 23°C) for approximately 12 h prior to microbial analysis to ensure complete thawing of the low-moisture feed samples before microbial processing. Because the samples remained dry during thawing, microbial growth during this period was considered minimal.

Table 1.

Number of feed samples collected from each feed mill facility.

Feed Mill	Ingredient	Starter	Grower	Finisher	Withdrawal	Breeder	Pullet	Total
A	0	7	6	0	10	0	0	23
B	6	6	6	0	12	0	0	30
C	9	4	3	0	6	0	0	22
D	9	3	3	0	6	0	0	21
E	5	1	1	1	0	1	0	9
F	5	1	1	1	0	1	0	9
G	0	2	2	2	2	2	2	12
H	2	2	2	0	2	0	0	8
I	2	2	0	0	0	4	0	8
J	0	2	2	2	2	0	2	10
K	0	2	2	2	2	2	2	12
L	3	1	1	0	2	0	0	7
M	4	1	1	0	2	0	0	8
N	0	1	0	0	0	0	0	1
Total	45	35	30	8	46	10	6	180

*This table summarizes the feed samples collected during the study by feed mill and sample type. Feed Mill N contributed a single sample because only one sample was available for collection during the study period.

Microbial analysis

For each sample, 2 g of feed was weighed and poured into a 7-oz filtered Whirl-Pak® bag with 10 mL of BPW. The filtered Whirl Pak® bag was then homogenized in a stomacher for one minute. The homogenized samples were then incubated for 24 h at 37°C. Microbiological sample processing and pathogen detection procedures were based on modified U.S. Food and Drug Administration Bacteriological Analytical Manual (FDA-BAM) methods, where appropriate (U.S. Food and Drug Administration, 2026).

Total count rapid, coliform, and E. coli isolation

Three types of Compact Dry plates (Shimadzu Diagnostics Corporation, 3-24-6, Ueno, Taito-ku, Tokyo 110-0005, Japan) were used to specifically detect TCR, Coliforms, and E. coli. Following Compact Dry’s general testing protocol, 1 mL of each sample was dispensed onto each plate, and the lid was replaced to allow the sample to diffuse onto the dehydrated media sheet. Each sample was diluted by 10⁻¹, 10⁻², and 10⁻³ for the three types of Compact Dry plates. Plates were incubated for 24 h at 37°C, after which colonies were counted to determine prevalence. TCR aerobic colonies were red in color, Coliforms were blue to blue-green colonies, and E. coli were identified as blue to blue-purple colonies.

Salmonella isolation

Prevalence of Salmonella was determined using modified FDA-BAM methods (U.S. Food and Drug Administration, 2026). The homogenized feed and BPW samples were incubated for 24 h at 37°C. One mL of each homogenized feed sample was aliquoted into 9 mL of Tetrathionate (TT) broth and 9 mL of Rappaport-Vassiliadis (RV) broth, and the mixture was vortexed. The TT broth and RV broth were then incubated for 24 h at 37°C. Following the incubation period, 100 µL of each broth sample was spread-plated onto XLT-4 agar and incubated for 24 h at 37°C. The XLT-4 plates were observed, and all plates that grew black colonies were recorded as presumptive-positive. All presumptive-positive plates were individually sealed with parafilm and stored at 0°C in a refrigerator until DNA extraction and PCR analysis.

Campylobacter isolation

Prevalence of Campylobacter was determined using a modified enrichment and isolation procedure adapted from ISO 10272-1 (International Organization for Standardization, 2017). The homogenized feed and BPW samples were incubated for 24 h at 37°C. One mL of each homogenized feed sample was aliquoted into 9 mL of 2x Bolton’s Broth (BB), vortexed, and incubated at 42°C for 48 h in a microaerophilic jar under microaerophilic conditions (5% O₂, 10% CO₂, and 85% N₂). Following incubation, 100 µL of the enriched BB sample was spread-plated onto Campylobacter Agar (Oxoid™ Campylobacter Agar Base, Thermo Fisher Scientific, Madison, WI, USA) and incubated at 42°C for 48 h under the same microaerophilic conditions. The Campylobacter plates were observed, and all plates that grew clear/light grey colonies were recorded as presumptive-positive. All presumptive-positive plates were individually sealed with parafilm and stored at 0°C in a refrigerator until DNA extraction and PCR analysis.

DNA extraction

A single colony from each Salmonella and Campylobacter presumptive-positive plate was added to a 10 mL tube of Tryptic Soy Broth (TSB) and incubated for 24 h at 37°C. The TSB samples were then pelleted by centrifugation in an Eppendorf 5810 R centrifuge (Eppendorf, Hamburg, Germany) at 4,000 rpm for 6 minutes. To wash the pellet, the TSB supernatant was poured off, 1 mL of Phosphate Buffered Saline (PBS; HyClone™, Cytiva, Logan, UT, USA) was added to resuspend the pellet, and the mixture was centrifuged at 4,000 rpm for 6 minutes. All samples were washed three times before DNA extraction. Extraction of Salmonella DNA and Campylobacter DNA was performed by using the ZYMO Research Quick-DNA™ Fecal/Soil Microbe Miniprep Kit (ZYMO Research, Irvine, CA, USA) following the manufacturer’s protocol.

Following DNA Extraction, Nanodrop One (Thermo Fisher Scientific, Madison, WI, USA) was used to determine the concentration and purity (260/280) of each sample. Nanodrop analysis was used to assess DNA degradation and confirm the presence/absence of DNA before PCR. Gel electrophoresis was run. To prepare samples, a dilution calculator (70 ng/µL desired concentration, 5 µL final solution volume) was used to determine µL of DNA and molecular H₂O needed to standardize DNA samples. For each sample, 5 µL of standardized DNA and 1 µL of 6x gel stain were mixed and added to the wells. All gels were run at 80 V for 45 min and were imaged using the Bio-Rad ChemiDoc Go Imaging System (Bio-Rad Laboratories, Hercules, CA, USA).

Salmonella polymerase chain reaction (PCR) confirmation

The traditional PCR procedure for Salmonella isolate confirmation followed Thames et al's. (2022) procedure. The invA gene was used as the amplified target gene using a PCR Thermo Cycler (Model 5435 Mastercycler epgradient S, Eppendorf, Hamburg, Germany). Salmonella-specific primers, S139 and S141 (Eurofins Genomics LLC, Louisville, KY), which were used as described by Rahn et al. (1992), consisted of the following nucleotide sequences: 5′-GTG AAA TTA TCG CCA CGT TCG GGC AA-3′ and 5′-TCA TCG CAC CGT CAA AGG AAC C-3′. The PCR reaction mixture consisted of 1 µL of DNA and 9 µL of master mix containing 5 µL of GoTaq Green (Promega, Madison, WI, USA), 0.25 µL of each forward and reverse primer, and 3.5 µL of molecular-grade water. Positive control was obtained from confirmed Salmonella isolates. Molecular water was used as a negative control. The PCR cycle conditions were as follows: an initial denaturation step at 94°C for 2 min, followed by 35 cycles of denaturation at 94°C for 30 s, annealing at 60°C for 30 s, and primer extension at 72°C for 30 s. A final incubation cycle was set at 72°C for 5 min.

Campylobacter PCR confirmation

Traditional PCR was used to amplify the targeted 23S rRNA gene using a PCR Thermo Cycler (Model 5435 Mastercycler epgradient S, Eppendorf, Hamburg, Germany). Campylobacter 23SF and 23 specific primers, which were used as described by Wang et al. (2002), have the following nucleotide sequences 5′-TAT ACC GGT AAG GAG TGC TGGAG-3′ and 5′-ATC AAT TAA CCT TCG AGC ACCG-3′. The positive control was obtained from confirmed Campylobacter isolates. Molecular water was used as a negative control. The PCR reaction mixture consisted of 1 µL of DNA and 9 µL of master mix containing 5 µL of GoTaq Green (Promega, Madison, WI, USA), 0.25 µL of each forward and reverse primer, and 3.5 µL of molecular-grade water. The PCR cycle conditions were as follows: an initial denaturation step at 95°C for 2 min, followed by 35 cycles of denaturation at 95°C for 30 s, annealing at 58°C for 30 s, and primer extension at 72°C for 1 min. A final incubation cycle was set at 72°C for 5 min.

PCR analysis

To confirm the amplification of the target gene, PCR products (284 bp DNA fragments) were electrophoresed on a 1% agarose gel containing SYBR™ Safe DNA gel stain (Invitrogen™, Carlsbad, CA, USA) in 1X Tris-Acetate-EDTA (TAE) buffer. For each gel, wells were loaded with a DNA ladder, positive control, negative control, and the corresponding samples. The PCR product (1 µL) was loaded into each well of the gel, and 1.5 µL of 1 kb PLUS™ DNA ladder (GoldBio, St. Louis, MO, USA) was used as a marker for the PCR products. An 80 V current was applied to each gel for 1 hour and 10 minutes. After running gel electrophoresis, the products were visualized using the Bio-Rad ChemiDoc Go Imaging System (Bio-Rad Laboratories, Hercules, CA, USA).

Statistical analysis

Feed and ingredient samples were collected from multiple integrated and commercial feed mills in 2024 and 2025 as part of a longitudinal observational study with repeated measures. Samples were analyzed for pathogen prevalence, with positive Salmonella and Campylobacter samples confirmed by PCR. Data were analyzed using the PROC GLIMMIX procedure of SAS version 9.4 (SAS Institute Inc., Cary, NC, USA) with a logistic regression model and a binomial distribution. Pairwise comparisons of least squares means were performed when appropriate, and statistical significance was declared at P ≤ 0.05.

Results

Compact dry results

Escherichia coli was detected in all ingredients and finished feed samples plated on Compact Dry (EC, TCR, and CF), regardless of feed mill location or feed type. Because all samples demonstrated growth at one or more dilutions, statistical comparisons among locations or feed types were not conducted.

Feed mill facility: RV and TT broth salmonella prevalence

Feed mill Salmonella prevalence data for each enrichment medium is shown in Table 2. No significant differences were observed between the feed mill facilities with either RV Broth (P=0.9330) and TT Broth (P=0.7991).

Table 2.

Salmonella feed mill facility RV and TT broth prevalence.

Feed mill facility	# of samples/facility	RV Broth Salmonella Prevalence %	TT Broth Salmonella Prevalence %
A	23	0.00	0.00
B	30	26.67	36.67
C	22	0.00	0.00
D	21	9.52	9.52
E	9	0.00	0.00
F	9	33.33	44.44
G	12	41.67	16.67
H	8	0.00	0.00
I	8	0.00	12.50
J	10	10.00	20.00
K	12	0.00	0.00
L	7	28.57	14.29
M	8	0.00	0.00
N	1	0.00	0.00
P-Value		0.9330	0.7991

*The values provide a comparison of detection outcomes between the two enrichment methods, highlighting potential differences in Salmonella recovery based on broth type and feed mill facility.

Feed mills A, C, E, H, I, K, M, and N did not recover Salmonella from RV Broth, and of these feed mills, only feed mill I recovered Salmonella from TT Broth. Feed Mill B and I were collected at the same feed mill location, but Feed Mill B (30 samples) was collected in 2024, and Feed Mill I (8 samples) was collected in 2025, with much fewer samples than the 2024 collection amount. Feed Mill C and M were collected from the same feed mill at two different times, suggesting that the feed mill employs effective control measures for Salmonella. Feed mills E, H, K, M, and N were collected from different locations, and only feed mill E was collected in 2024. PCR analysis was performed on all black-colony samples, and all were confirmed positive.

Feed type: RV and TT broth salmonella prevalence

Feed type Salmonella prevalence data for each enrichment medium are shown in Table 3. No significant differences were observed between sample types when enriched in RV Broth (P=0.1187) or TT Broth (P=0.6365).

Table 3.

Salmonella sample type RV and TT broth prevalence.

Sample Type	# of samples	RV Broth Salmonella Prevalence %	TT Broth Salmonella Prevalence %
Ingredient (I)	45	17.78	20.00
Starter (S)	35	11.43	14.29
Grower (G)	30	6.67	10.00
Finisher (F)	8	0.00	0.00
Breeder (B)	10	20.00	0.00
Pullet (P)	6	50.00	33.33
Withdrawal (W)	46	4.36	8.70
P-Value		0.1187	0.6365

*The values provide a comparison of detection outcomes between the two enrichment methods, highlighting potential differences in Salmonella recovery based on broth type and feed sample type.

The most collected sample types from each category were Ingredients (45 samples) and Withdrawal (46 samples). The pullet developer samples had the highest Salmonella prevalence in RV Broth (50.00%) and TT Broth (33.33%); however, this was based on only 6 samples. When comparing Salmonella prevalence to the number of samples per category, ingredient samples accounted for the highest proportion of Salmonella-positive samples. Of the 45 samples received, RV Broth recovered Salmonella from 8 samples (17.78%), and TT Broth recovered Salmonella from 9 samples (20.00%). PCR analysis was performed on all samples with black colonies, and all were confirmed positive.

Comparison of salmonella prevalence in feed samples by sample year

The comparison of Salmonella prevalence by sample year is presented in Table 4. No significant differences were observed between sample years for both RV Broth (P=0.0998) and TT Broth (P=0.1726).

Table 4.

Comparison of Salmonella prevalence in feed samples by sample year.

Year	# of samples/year	# of positive RV Broth Samples	RV Broth Salmonella Prevalence %	# of positive TT Broth Samples	TT Broth Salmonella Prevalence %
2024	106	10	9.43	13	12.26
2025	74	11	14.86	10	13.51
Total	180	21	11.67	23	12.78
P-Value			0.0998		0.1726

Salmonella samples were analyzed by two enrichment broths (RV and TT). In the 2024 samples, Salmonella prevalence in RV broth was 9.43%; in the 2025 samples, it increased to 14.86% (Table 4). Tetrathionate broth samples showed a similar trend: in 2024, Salmonella prevalence was 12.26%, and in 2025, it increased to 13.51% (Table 4). Following enrichment culture, PCR analysis was performed on all samples that grew black colonies, confirming that all were positive for Salmonella.

Feed mill facility: campylobacter BB before and after PCR analysis

The prevalence of Campylobacter in the feed mills before and after PCR is presented in Table 5. No significant differences were observed between the feed mill facilities before (P=0.8875) and after (P=0.9999) PCR confirmation.

Table 5.

Campylobacter feed mill facility prevalence before and after PCR analysis.

Feed Mill	Total # of positive samples before PCR	Total % of positive samples before	Total # of positive samples post-PCR	Total % of positive samples post-PCR
A	7	30.43	0	0.00
B	7	23.33	5	16.67
C	6	27.27	2	9.09
D	2	9.52	0	0.00
E	1	11.11	0	0.00
F	4	44.44	1	11.11
G	3	25.00	1	8.33
H	0	0.00	0	0.00
I	3	37.50	0	0.00
J	2	20.00	2	20.00
K	3	25.00	1	8.33
L	1	14.29	1	14.29
M	3	37.50	1	12.50
N	0	0.00	0	0.00
P-Value		0.8875		0.9999

*The values compare presumptive-positive and PCR-confirmed Campylobacter detections among feed mill facilities.

Following enrichment culture, PCR analysis was performed on all samples that grew light grey colonies to confirm Campylobacter positivity. Feed mills A, B, C, D, E, F, G, I, J, K, L, and M yielded presumptive-positive cultures (42 positive). After PCR, only feed mills B, C, F, G, J, K, L, and M were confirmed to be Campylobacter (14 positive). No positive colonies were confirmed by PCR from Feed Mills A, D, E, and I. No positive Campylobacter colonies were detected from Feed Mills H and N.

Feed sample: type campylobacter BB before and after PCR analysis

Prevalence of Campylobacter in the sample types before and after PCR is presented in Table 6. No differences were observed between feed types pre (P=0.2905) and post (P=0.1002) PCR confirmation.

Table 6.

Campylobacter sample type prevalence before and after PCR analysis.

Feed Type	Total # of positive samples before PCR	Total % of positive samples before	Total # of positive samples post-PCR	Total % of positive samples post-PCR
Ingredient (I)	8	17.78	5	11.11
Starter (S)	12	34.29	2	5.71
Grower (G)	9	30.00	1	3.33
Finisher (F)	0	0.00	0	0.00
Breeder (B)	3	30.00	1	10.00
Pullet (P)	3	50.00	3	50.00
Withdrawal (W)	7	15.22	2	4.35
P-Value		0.2905		0.1002

*The values compare presumptive-positive and PCR-confirmed Campylobacter detections among feed sample types.

Following enrichment culture, PCR analysis was performed on all samples that grew light grey colonies to confirm Campylobacter positivity. Starter feed samples represented the most presumptive-positive Campylobacter samples before and after PCR. Ingredient and grower feed had 8 and 9 presumptive-positive samples, respectively, before PCR confirmation; however, only 5 ingredient samples and 1 grower sample were confirmed by PCR. All Campylobacter-positive pullet samples were confirmed positive after running PCR. No Campylobacter was detected in the Finisher feed category. In total, prior to PCR, 42 samples were suspected to be Campylobacter-positive (23.33%), but PCR confirmed that only 14 samples (7.78%) were Campylobacter-positive.

Comparison of campylobacter prevalence in feed samples by sample year

The comparison of Campylobacter prevalence by sample year is presented in Table 7. No differences were observed between the pre-PCR (P=0.1165) and post-PCR (P=0.1714) confirmation samples.

Table 7.

Comparison of Campylobacter prevalence in feed samples by sample year.

		Prior to PCR Analysis		Post-PCR Analysis
Year	# of samples/year	# of positive BB Broth Samples	BB Broth Campylobacter Prevalence %	# of positive BB Broth Samples	BB Broth Campylobacter Prevalence %
2024	106	23	21.70	7	6.60
2025	74	19	25.68	7	9.46
Total	180	42	23.33	14	7.78
P-Value			0.1165		0.1714

Campylobacter samples were analyzed by plating enriched BB, and light grey colonies were presumptive positives. In the 2024 samples, 21.70% were presumed positive prior to PCR. In the 2025 samples, 25.68% were presumed positive prior to PCR. After PCR confirmation, 6.60% of the samples from 2024 and 9.46% of the samples from 2025 were confirmed positive for Campylobacter.

Discussion

Microbial contamination of feed samples by feed type

All feed samples collected were from an assortment of raw ingredients, starter, grower, finisher, pullet developer, breeder, and withdrawal feed. Historically, raw ingredients and cereal grains have been considered the primary vehicles for microorganisms entering the feed mill. Previous research indicates that the receiving operations area of the feed mill is where the highest levels of microorganisms are detected, as cereal grains are at the highest risk of carrying microorganisms (Jones, 2011; Davies and Wray, 1997). Cereal grains may become contaminated with microorganisms from the soil they were grown in, during storage or before shipping, and during transportation to the feed mill. Parker et al. (2019) analyzed 22 Australian feed mills from January 2003 to May 2018 and found that the highest Salmonella prevalence was in raw feed components. Apart from the pullet developer feed samples, our study aligns with Parker et al. and demonstrates that raw ingredients had the highest Salmonella prevalence among starter, grower, finisher, and withdrawal feed samples. These results highlight the need to further investigate each pathogen’s persistence and survivability throughout feed production, as well as its ability to contaminate equipment, potentially creating a source of microbial cross-contamination.

Thermal treatment variations by feed mills

Our findings, which show the prevalence of microorganisms across nearly all feed samples tested, emphasize the importance of implementing measures to manage contamination and cross-contamination throughout the feed manufacturing process. In most feed mills, thermal treatment is part of the feed manufacturing process, which involves conditioning, retention, and pelleting as a “kill step” to eliminate microorganisms in feed. Time and temperature parameters fluctuate during conditioning depending on the feed mill’s equipment and geographical location.

Not all feed mills included in this study provided conditioning times and temperatures. Feed Mills C and M originated from the same feed mill but represented two different sampling periods and reported a conditioning temperature of 85°C for 90 seconds. Interestingly, this feed mill did not recover any Salmonella from RV or TT broth enrichment methods; however, 5 feed samples tested positive for Campylobacter after PCR analysis. To date, no published studies have evaluated thermal inactivation of Campylobacter in feed. Previous research conducted by Coe et al. demonstrated the differences of thermal inactivation when comparing Salmonella Typhimurium to a surrogate Enterococcus Faecium (Coe et al., 2022). These findings emphasize the need to further investigate the survivability of various microorganisms compared to standard conditioner times in integrated feed mills. One limitation of the current study is that feed samples were initially pre-enriched in buffered peptone water under aerobic conditions prior to selective enrichment in Bolton's broth. Because Campylobacter spp. are microaerophilic organisms, aerobic pre-enrichment may have reduced the recovery of viable cells, contributing to an underestimation of Campylobacter prevalence. Future studies evaluating Campylobacter contamination in feed manufacturing environments should consider direct enrichment in Bolton's broth or tryptic soy broth under microaerophilic conditions to improve organism recovery and detection sensitivity.

Recontamination, post-thermal treatment

All feed types collected after heat treatment exhibited substantially lower, though not eliminated, prevalence of Salmonella and Campylobacter compared to the Ingredient and Starter categories. This is likely to be due to potential recontamination in the coolers or storage. Previous research highlighted coolers as a post-heat-treatment contamination “hot spot” for microbial growth, as they provide favorable, warm, moist, and humid environments for microbes to thrive and replicate (Davies et al., 2001; Gosling et al., 2022). Even microorganism-free feed in coolers remains at risk of recontamination due to the high airflow and the difficulty of cleaning the equipment (Gosling et al., 2022).

Globally, there is an abundance of research evaluating pathogen presence in feed manufacturing facilities; most studies focus on the finished feed product, and there is a gap in the literature on pathogen detection within the feed mill environment. The feed mills participating in this current study did not provide specific sampling sites; it is hypothesized that most finished feed samples were taken during the post-cooling stage of feed production. While this is an effective sampling location for feed manufacturers, as it represents the final exit point of the feed mill, there is a risk of false negatives due to pathogens’ tendency to cluster and their lack of distribution within the feed. Muñoz et al. (2021) evaluated the presence of Salmonella, E. coli, and Clostridium perfringens from 4 commercial feed mills and a university feed mill from the following sample locations: ingredient receiving, post mixing, post pelleting (cooling), and at loadout. Interestingly, Salmonella was not detected in any feed ingredient or processed feed sample, but E. coli and Clostridium perfringens were detected in processed feeds at various feed mill locations (Muñoz, 2021). Collectively, the results suggest that pathogen detection accuracy is maximized when multiple sites are sampled, highlighting the limitations of relying solely on finished feed testing.

Microbial contamination of commercial vs integrated feed mills

Both feed mill types, commercial and integrated, demonstrated microbial presence within all sample types. The 22 feed mills evaluated in the study by Parker and coauthors (2019) were commercial feed mills that produced ruminant, poultry, and swine diets. In general, commercial feed mills have a higher risk of contamination and harbor more diverse microbial populations because they acquire ingredients from multiple suppliers and store finished feed for longer (usually 1–2 weeks). Results from our study demonstrated that both commercial and integrated feed mill categories included facilities with 0% prevalence of Salmonella and Campylobacter. Because sample numbers varied among participating feed mills and several facilities had no positive detections, statistical power to detect differences among feed mills may have been limited. To our knowledge, limited data have been published on microbial contamination across different types of feed mills. Addressing this gap through future studies is important to better understand the frequency of occurrence and variability of microbial pathogens in these facilities.

Limitations to feed sample further analysis

Historically, conventional culture methods have been used as the gold standard for detecting pathogens in poultry feed (Cason et al., 2025). All feed samples that were plated using the conventional culture method and were presumed Salmonella positive were confirmed positive after conducting PCR. While these methods are well established, they only detect live organisms capable of growth and take a lengthy time (4–5 days) to confirm a result. Conventional PCR is widely used as a confirmatory diagnostic tool; however, because it amplifies DNA from both live and dead microorganisms, positive PCR results should be interpreted with caution when assessing food safety risks, as DNA detection alone does not necessarily indicate the presence of viable organisms capable of causing infection. In this study, 4 feed mills that were presumed to be Campylobacter-positive prior to PCR testing had 0% Campylobacter prevalence post-PCR. Therefore, our study suggests that additional PCR confirmation is necessary to further evaluate pathogens in feed, as some may be detected at low levels, potentially leading to false negatives or a viable but nonculturable state.

Research on conventional culture methods evaluating feed ecology and identifying microorganisms is well established; however, applying microbiome analyses to feed is hindered by the complexity of feed microbial communities. The term “complete feed” describes the complexity of the feed, which includes various cereal grains, crystalline amino acid supplements, micronutrients, and other specialized ingredients (Olson et al., 2022). Although previous research identified cereal grains as the main source, each component can serve as a vector for microorganisms in feed.

At present, there is minimal research evaluating microbiome analyses of feeds, partially because microbiome feed analysis is a relatively recent opportunity and partially because of the previously mentioned challenges that remain unsolved. Therefore, this study evaluated presumptively positive Salmonella and Campylobacter colonies by performing DNA extraction and PCR to verify their positivity. Currently, there is no distinct DNA extraction kit for animal feed due to the complex feed matrix. As previously mentioned, the ZYMO Research Quick-DNA^TM Fecal/Soil Microbe Kit was used for DNA extraction, as this kit can extract PCR-quality DNA from gram-positive and gram-negative microbes. A study by Wegi et al. examined intestinal contents and feces from poultry and swine and evaluated multiple DNA extraction methodologies (Wegi et al., 2021). The results of this study demonstrated that the MP Biomedicals FastDNA^TM Spin kit (FSP) and the Zymo Quick-DNATM Fecal/Soil Microbe Kit (QDK) yielded the highest DNA yields (Wegi et al., 2021).

The results of the current study demonstrate that effective feed mill management practices, including heat treatment, biosecurity, and sanitation, can substantially reduce the risk of pathogen contamination, such as Salmonella and Campylobacter, in integrated and commercial feed manufacturing. However, the widespread recovery of coliforms across all facilities and sample types highlights the persistence of background microbial contamination within feed mill environments and indicates opportunities for improved decontamination strategies beyond thermal processing. Elevated microbial levels in ingredients and starter feeds further support the role of heat treatment in reducing microbial load, while also suggesting potential post-processing recontamination. Targeted environmental monitoring, particularly aerosol sampling in post-heat-treatment areas, may provide critical insights into contamination pathways and support the development of more effective feed mill hygiene interventions.

CRediT authorship contribution statement

Emily Magee: Writing – original draft, Formal analysis. Li Zhang: Writing – review & editing, Methodology. Cangliang Shen: Writing – review & editing, Methodology, Funding acquisition. Timothy Boltz: Writing – review & editing, Writing – original draft, Investigation, Funding acquisition, Conceptualization.

Disclosures

The authors declare that they have no conflict of interest.