Hostname: page-component-7b7567c8c-kwt9d Total loading time: 0 Render date: 2026-11-12T09:23:24.849Z Has data issue: false hasContentIssue false
  • English
  • Français

Risk factors of Shiga toxin-producing Escherichia coli in livestock raised on diversified small-scale farms in California

Published online by Cambridge University Press:  01 June 2022

Laura Patterson,
Nora Navarro-Gonzalez,
Michele T. Jay-Russell,
Peiman Aminabadi and
Alda F. A. PiresOpen the ORCID record for Alda F. A. Pires[Opens in a new window]
Laura Patterson
Affiliation:
Department of Population Health and Reproduction, University of California-Davis, Davis, CA 95616, USA
Nora Navarro-Gonzalez
Affiliation:
Western Center for Food Safety, University of California-Davis, Davis, CA 95616, USA
Michele T. Jay-Russell
Affiliation:
Western Center for Food Safety, University of California-Davis, Davis, CA 95616, USA
Peiman Aminabadi
Affiliation:
Western Center for Food Safety, University of California-Davis, Davis, CA 95616, USA
Alda F. A. Pires*
Affiliation:
Department of Population Health and Reproduction, University of California-Davis, Davis, CA 95616, USA
*
Author for correspondence: Alda F. A. Pires, E-mail: apires@ucdavis.edu
Rights & Permissions [Opens in a new window]

Abstract

The increasing number of diversified small-scale farms (DSSF) that raise outdoor-based livestock in the USA reflects growing consumer demand for sustainably produced food. Diversified farms are small scale and raise a combination of multiple livestock species and numerous produce varieties. This 2015–2016 cross-sectional study aimed to describe the unique characteristics of DSSF in California, estimate the prevalence of Shiga toxin-producing Escherichia coli (STEC) in livestock and evaluate the association between risk factors and the presence of STEC in livestock, using generalised linear mixed models. STEC prevalence was 13.62% (76/558). Significant variables in the mixed-effect logistic regression model included daily maximum temperature (OR 0.95; CI95% 0.91–0.98), livestock sample source (cattle (OR 4.61; CI95% 1.64–12.96) and sheep (OR 5.29; CI95% 1.80–15.51)), multiple species sharing the same barn (OR 6.23; CI95% 1.84–21.15) and livestock having contact with wild areas (OR 3.63; CI95% 1.37–9.62). Identification of STEC serogroups of public health concern (e.g. O157:H7, O26, O103) in this study indicated the need for mitigation strategies to ensure food safety by evaluating risk factors and management practices that contribute to the spread and prevalence of foodborne pathogens in a pre-harvest environment on DSSF.

Keywords

Food safetymixed crop–livestock farmsorganic productionpasture-raised livestockShiga toxin-producing Escherichia colisustainably produced

Information

Type
Original Paper
Information
Epidemiology & Infection , Volume 150 , 2022, e125
Check for updates
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
Copyright © The Author(s), 2022. Published by Cambridge University Press

Introduction

The increasing number of diversified small-scale farms (DSSF) that raise outdoor-based (i.e. grass-fed, pasture-raised) livestock in the USA reflects growing consumer demand for sustainably produced local foods, including animal products such as meat and eggs [Reference Hilimire1–Reference Kremen, Iles and Bacon5]. California is the top producer in the USA of agricultural products, as well as organic food sales, which includes products from DSSF [3, Reference Cesca6]. However, there is a lack of science-based information characterizing the risk factors associated with the prevalence of foodborne pathogens, such as Shiga toxin-producing Escherichia coli (STEC), in livestock raised on DSSF.

Diversified farms are most often small scale and can raise a combination of multiple livestock species and numerous produce varieties, with the intent of selling specialised (e.g. organically grown) animal products directly to consumers [Reference Kremen, Iles and Bacon5]. Some DSSF integrate livestock and crop production (i.e. integrated crop–livestock farms) by using their animals to graze crop residues or cover crops before planting a field with fresh produce [Reference Hilimire1, Reference Salaheen7]. Grazing improves soil fertility and provides farm owners with another source of revenue through fibre or meat products [Reference Hilimire1, Reference Hoar8]. Additionally, many countries promote diversified farming systems as an alternative to intensive agriculture, displaying a willingness to transition to agroecology, in line with the UN Sustainable Development Goal number 12: ‘Ensure sustainable consumption and production patterns’ [9]. This goal may lead to an increase in the number of DSSF and products globally.

Many consumers perceive produce and meat raised on small-scale farms as more natural or safer (e.g. pesticide-free) than food grown on large-scale conventional farms or meat animals raised in confinement systems [Reference Martinez2, Reference Harvey, Zakhour and Gould10]. However, livestock are asymptomatic reservoirs for foodborne pathogens, and without adequate mitigation strategies, these pathogenic microorganisms may enter the food supply [Reference Pires11–Reference Jacob16]. Livestock are intermittent shedders of enteric pathogens and shedding may increase under certain conditions, such as during specific seasons or periods of stress (e.g. transportation), or due to certain management practices (e.g. stocking rate) [Reference Salaheen7, Reference Doyle and Erickson17–Reference Moriarty19]. STEC can survive in the soil for extended periods of time and spread to humans directly through contact with livestock or indirectly via contaminated food (e.g. fresh produce, meat) or water [Reference Ongeng15, Reference Doyle and Erickson17, 20–Reference Mughini-Gras22]. STEC remains one of the main enteric pathogens associated with foodborne outbreaks in the USA [20, Reference Conrad21]. The top seven STEC O-serogroups that cause the most severe illness in humans include O157:H7, O26, O103, O111, O121, O145 and O45 [23]. Beef products and vegetables are common sources of infection [20,Reference Conrad21]. However, outbreaks linked to organic animal products have been reported in small numbers [Reference Harvey, Zakhour and Gould10]. Fresh produce consumed raw are considered high-risk crops, and when contaminated by livestock or wildlife faeces containing STEC, they can become a vehicle for these pathogens to enter the food supply [Reference Conrad21,Reference Mughini-Gras22, Reference Strawn24]. For instance, cattle grazing uphill from a produce field was likely the causative factor for the 2019 E. coli O157:H7 romaine lettuce outbreak [25].

The aim of this study was to (a) describe the unique characteristics of DSSF, (b) estimate the prevalence of STEC in livestock and (c) evaluate the association between risk factors and the presence of STEC in livestock raised on DSSF located in California.

Methods

Study design and farm enrolment

During 2015–2016, we conducted a repeated cross-sectional study in California to test for Campylobacter spp., Salmonella spp., non-O157 STEC and E. coli O157:H7 in faecal samples collected from livestock raised on DSSF [Reference Pires11]. Enrolment criteria for participating DSSF was based on the United States Department of Agriculture Economic Research Service (USDA-ERS) definition of small-scale farms [4], but adjusted to fit the unique growing conditions of California operations: (1) small- to medium-scale farm (i.e. poultry producers raising <1000 birds annually or livestock producers with an annual gross-sales <$500 000 and that raised a maximum of 500 goats/sheep and/or 100 cattle and/or 100 pigs); (2) raised a diversity of crops and/or multiple species of livestock and poultry; (3) marketed farm products directly to consumers (e.g. farmers markets, community-supported agriculture programs); and (4) willingness to participate. We enrolled farms from four regions of northern and central California that corresponded to the diversity of bioregions and that enabled an acceptable time between sample collection and laboratory processing [26]. Farm recruitment occurred through personal communications, announcements at agriculture outreach seminars and contact at farmers' markets or agricultural fairs.

Sample collection

Farms were visited twice between May 2015 and June 2016, once during each of the following periods: summer/autumn or winter/spring, which reflect California's growing seasons and the seasonality of STEC shedding [Reference Pires11]. Livestock species sampled in this study included dairy and beef cattle, dairy and meat goats, pigs and sheep. Sample sizes were calculated using Epitools [27] based on the total number of animals on each farm, with an assumed STEC prevalence of 5% and 10% precision error, stratified by each livestock species. Individual fresh faecal samples were collected from the ground of barns or pastures. Samplers wore gloves and placed approximately 50–100 g of faeces into each sterile whirl-pak bag (Nasco, Modesto, CA, USA). Bags were immediately placed into plastic coolers containing ice packs, transported to the laboratory at the end of the sampling day and most samples were processed within 24 h.

Escherichia coli O157:H7 and non-O157 STEC culture and PCR

STEC was isolated from faecal samples as described previously with modifications [Reference Cooley28]. In brief, 10 g of faecal material was placed in 90 ml of Tryptic Soy Broth (TSB; Becton, Dickinson and Company, Franklin Lakes, NJ, USA). Samples were then incubated for 2 h at 25 °C with 100 rpm agitation, followed by 8 h at 42 °C with 100 rpm agitation and held overnight at 6 °C, using a Multitron programmable shaking incubator (Eppendorf, Hauppauge, NY, USA). For detecting E. coli O157:H7, immunomagnetic separation using Dynal anti-E. coli O157 beads (Invitrogen/Dynal, Carlsbad, CV, USA) was performed on TSB enrichment broths with the automated Dynal Bead Retriever (Invitrogen) per the manufacturer's instructions. After incubation and washing, 50 μl of the resuspended beads were plated onto Rainbow agar O157 (Biolog, Hayward, CA, USA) with novobiocin (20 mg/l) and tellurite (0.8 mg/l) (MP Biomedicals, Solon, OH, USA). Fifty microlitres of the resuspended beads were also plated onto MacConkey II Agar using sorbitol supplemented with potassium tellurite (2.5 mg/l) and Cefixime (0.05 mg/l) (CT-SMAC); plates were streaked for isolation and incubated for 24 h at 37 °C. Suspect E. coli O157:H7 isolates were confirmed using traditional PCR for the rfbE gene [Reference Philpott and Ebel29].

To detect non-O157 STEC, 1 ml of TSB enrichment broth was incubated in mEHEC selective media (Biocontrol, Bellevue, WA, USA) for 12 h at 42 °C followed by plating and incubating on Chromagar STEC (DRG International Inc., Springfield, NJ, USA) at 37 °C overnight. Up to eight presumptive STEC-positive colonies, demonstrating mauve colour and glowing under UV light, were confirmed for the presence of stx1 and/or stx2 genes by real-time PCR [Reference Cooley28]. Confirmed STEC isolates were then characterised for virulence genes (stx1, stx2, eaeA, hlyA and ehxA) using conventional PCR [Reference Philpott and Ebel29]. After PCR testing, one isolate with a unique virulence gene profile from each positive sample was submitted to the Pennsylvania State University E. coli Reference Center to confirm O-serogroups [Reference Cooley28, Reference DebRoy, Roberts and Fratamico30].

Farmer questionnaire and environmental factors

A 41-question questionnaire, consisting of mainly closed-ended questions, was given to farm owners mostly in person. The questionnaire included sections regarding farm demographics, animal health and farm management practices. Variables analysed for model building included risk factors from the farmer questionnaire, sample day factors (e.g. temperature) and farm-level characteristics (e.g. whether a farm raised multiple types of livestock). Weather data from the nearest California Irrigation Management Information System weather station (http://www.cimis.water.ca.gov) within a similar microclimate provided environmental factors to be included in model building (e.g. average relative humidity, daily maximum temperature) [31]. Also, USDA plant hardiness zones, which are based on the average annual minimum winter temperature, were included as a proxy for the many microclimates in California [32]. Only three zones were necessary to categorise our participating farms: zone 7b (−15 to −12.2 °C), 9a (−6.7 to −3.9 °C) and 9b (−3.9 to −1.1 °C).

Statistical data analysis and model building

Descriptive statistics (e.g. mean, range) were calculated for all data. STEC prevalence was estimated for the overall study and per livestock species. Generalised linear mixed models were used to assess the association between STEC presence in faecal samples and risk factors. The binary outcome of interest was whether each faecal sample was STEC-positive or negative. Univariate analysis assessed the distribution of variables. During bivariate analysis, variables with low variability, small cell sizes (<5) or large standard errors were either modified, collapsed if appropriate or discarded from model building. Correlation of numeric variables was measured with Spearman's rank correlation coefficient; those variables that were correlated 0.80 or more were highlighted during the model-building phase to evaluate for multicollinearity. To identify possible confounders, each variable was evaluated using a directed acyclic graph.

The glmer function was used from the lme4 package in R to build models, with farm added as a random effect to account for farm-level clustering effects [Reference Douglas, Martin and Ben Bolker33]. Manual two-way stepwise variable selection was employed for model building. Variance inflation factors (VIF) identified possible multicollinearity and variables with a VIF over five were assessed for removal. Top models were compared and a final model was chosen based on the lowest Akaike Information Criterion (AIC) and smallest deviance, relative to the other models. Intraclass correlation (ICC) was calculated, which accounts for the proportion of the model variance explained by clustering and indicated whether a random effect was necessary. Diagnostic plots from the DHARMa package in R were used to assess the final model and included fitted vs. binned residuals, a Q-Q plot and the Kolmogorov–Smirnov test statistic [Reference Hartig and Lohse34]. Odds ratios and 95% confidence intervals (CI95%) were calculated for variables in the final model. All data analyses were performed using R Statistical Software version 1.4.1036© [35].

Results

Study participants and farm characteristics

Sixteen farms participated in our study and were located in four regions of California: Central Valley (n = 10), Central Coast (n = 3), Shasta Cascade (n = 2) and the North Coast (n = 1). Four of the 16 farms (25.00%) raised livestock only, nine farms (56.25%) raised a combination of produce and livestock and used the animals to graze crop fields and three produce farms (18.75%) rented sheep seasonally to graze cover crops and did not raise any other animals. The rented sheep were present during the sampling day. Four farms were not organic (25.00%), 56.25% (9/16) were certified organic and 18.75% (3/16) used organic practices but were not certified by a third party.

STEC prevalence

A total of 558 faecal samples were collected from 16 farms. Overall STEC prevalence (i.e. O157:H7 and non-O157 STEC) was 13.62% (76/558; CI95% 10.88–16.75%). Farm-level STEC prevalence ranged from 0% to 30%; however, 37.50% (6/16) of farms had no positive samples. Of the 62.50% (10/16) of farms with positive STEC samples, the mean prevalence was 17.24%, with a median of 16.73%. Positive STEC samples were detected in all sampled livestock species. The count of each STEC O-serogroup and associated virulence genes per livestock species are shown in Table 1. Beef cattle had the highest STEC prevalence at 27.66% (13/47; CI95% 15.62–42.64%) and second highest were dairy cattle at 18.18% (12/66; CI95% 9.76–21.61%), with a 22.12% prevalence for all cattle combined. Goats had the next highest prevalence at 16.13% (15/93; CI95% 9.32–25.20%), and then sheep 13.40% (28/209; CI95% 9.09–18.78%). Swine samples had the lowest prevalence: 5.59% (8/143; CI95% 2.45–10.73%). Escherichia coli O157:H7 was found only in cattle, with a prevalence of 5.31% (6/113). Of the 76 positive STEC faecal samples, 72 were non-O157 STEC, six were E. coli O157:H7 and one was untypeable. Three samples from beef cattle were positive for both E. coli O157:H7 and a non-O157 STEC. All six O157:H7 positive isolates were from one farm and contained virulence factors stx2, ehxA, hlyA and eaeA, but not stx1. O26 was the most prevalent O-serogroup, accounting for 27.85% (22/79) of positive isolates. Stx2 was identified in 16.46% (13/79) of isolates, stx1 was identified in 69.62% (55/79) and 13.92% (11/79) contained both stx1 and stx2. In Table 1, isolates that shared the same virulence genes were from the same farm except the following, which originated from two farms each: O22 in dairy cattle, O103 in swine and O26 in swine. Additionally, O26 in goats was collected from three farms.

Table 1.

Virulence genes discovered in 79 isolates from 76 positive Shiga toxin-producing Escherichia coli samplesa collected from 16 California diversified small-scale farms, between May 2015 and June 2016, characterised by livestock species, O-serogroup and virulence gene

a One isolate was untypeable.

b Number of isolates containing the same O-serogroup and virulence genes.

Risk factor analysis

Of the 16 participating farms, two were excluded from model building, due to incomplete questionnaires, leaving a total of 502 samples for model building. The mean, median and range of selected continuous variables assessed for inclusion during model building are shown in Table 2. Farms in this study ranged from two to 500 acres and had been farming 2–30 years. Stocking density rate was calculated by dividing the total number of livestock species, excluding poultry, by the total number of farm acres. Selected categorical variables, stratified by whether they were STEC-positive or negative, are presented in Table 3, and P-values were reported for χ 2 test or Fisher's exact test if cell sizes were less than five. For instance, 28.99% (20/69) of positive samples came from farms that mixed livestock species within a barn, whilst only 15.70% (68/433) of negative samples came from farms with shared barns (P-value 0.012). Moreover, 72.46% (50/69) of positive samples were from farms that allowed livestock to have contact with bordering wild areas (e.g. streams, forest) (P-value 0.026).

Table 2.

Mean, median and range (i.e. minimum and maximum) of selected continuous variables assessed for model building and collected during a cross-sectional study conducted from 2015 to 2016 on 14a diversified small-scale farms in California

a Two of the 16 participant surveys were not completed.

Table 3.

Bivariate analysis for selected categorical variables stratified by positive (n = 69) or negative (n = 433) Shiga toxin-producing Escherichia coli status samples were collected during a cross-sectional study from 2015 to 2016 on 14c diversified small-scale farms in California

a USDA zone information can be found online: https://planthardiness.ars.usda.gov. Three USDA zones were used to categorise participating farms: zone 7b (−15 to −12.2 °C), 9a (−6.7 to −3.9 °C) and 9b (−3.9 to −1.1 °C).

b Organic: some farms use organic practices, but were ‘not certified’ organic by a third party.

c Two of the 16 farm surveys were not completed.

* Indicates statistical significance with a P-value < 0.05.

Final multivariable model results

The final mixed-effect multivariable logistic regression model is shown in Table 4 and had an AIC of 369.5, and a deviance of 351.5. The highest VIF for any variable in the model was only 2.02, which was below our threshold of five.

Table 4.

Association between risk factors and the presence of Shiga toxin-producing Escherichia coli in faecal samples (n = 502) collected from 14a diversified small-scale farms in California between May 2015 and June 2016, as demonstrated by a generalised linear mixed model, with farm as a random effect

a Two of the 16 participant surveys were not completed.

*Indicates statistical significance with a P-value < 0.05.

Significant risk factors associated with the presence of STEC in the final model included: for every increase in the daily maximum temperature °C, the odds of a STEC-positive sample decreased (OR 0.95; CI95% 0.91–0.98). The odds of a positive STEC sample were 4.61 times higher for cattle (OR 4.61; CI95% 1.64–12.96) compared to swine and more than five times greater in sheep (OR 5.29; CI95% 1.80–15.51). The odds of STEC increased by 6.23 times for those farms that housed multiple livestock species within the same barn vs. those farms that kept livestock separately (OR 6.23; CI95% 1.84–21.15). The odds of a positive STEC sample were more than three times greater (OR 3.63; CI95% 1.37–9.62) for a farm that allowed its livestock contact with wild areas. The effect of the number of years a farm had been in operation was not significant (P-value 0.076; CI95% 0.13–1.11), but was included in the model as a possible confounder.

Diagnostically, the simulated residuals vs. predicted values did not show any significant problems: the Q-Q plot from DHARMa simulated scaled residuals was linear with no major deviations, and the Kolmogorov–Smirnov test indicated no deviation from uniform distribution of the scaled residuals [Reference Hartig and Lohse34]. The adjusted ICC was 0.08 for the final model, which signified the proportion of the variance that was explained by farm clustering and indicated the need for a farm random effect. The isSingular test function in the lme4 package was false, which indicated no singularities existed in the final model.

Discussion

This is one of the first studies to describe the unique characteristics of DSSF in California while ascertaining significant associations between risk factors and the prevalence of STEC. This study detected STEC on more than half of the enrolled farms and typed O-serogroups including ones that cause serve illness in humans, including O157:H7, O26, O103 and O111 [23]. Significant risk factors for the presence of STEC included a negative association with the daily maximum temperature, whether multiple livestock species shared a barn, the host species of the collected faecal sample, and whether livestock had contact with wild areas.

Overall STEC prevalence measured for the 16 farms in this study was 13.62%. Six of the 16 farms had 0% STEC prevalence; however, due to the intermittent shedding of STEC which is affected by many factors, this result does not necessarily indicate that their livestock were not STEC reservoirs. Although STEC prevalence in livestock raised on large farms has been measured frequently in past studies, evaluation of STEC prevalence and associated risk factors estimated on DSSF is less common [Reference Cooley28, Reference Tseng and Fratamico36–Reference Tseng38]. However, a study conducted by USDA-APHIS collected faecal samples from dairy cows in 21 states and stratified E. coli O157:H7 prevalence between large dairies (i.e. 500 or more cows) and small dairies (i.e. 100 cows or less) and reported that small ranches had 29.4% E. coli O157:H7 and large dairies had 53.9% prevalence [39]. Although this report indicated that small farms have less E. coli O157:H7 than large farms, the 29.4% prevalence they detected on small dairies is still larger than the 18.18% we identified in dairy cattle. Risk factors for the transmission of foodborne pathogens on large farms may be different, especially if they only raise one crop or livestock type, instead of a diversity of species as seen on DSSF.

One of our previous studies measured a 4.17% STEC prevalence in sheep raised on a mixed crop–livestock organic farm in California, which was lower than the 13.4% prevalence in sheep identified in this study [Reference Patterson40]. A 2005 California study that screened livestock at the California state fair, which usually hosts livestock raised on small-scale operations, observed a 3% prevalence of E. coli O157:H7 in pigs, but did not find O157:H7 in any other livestock samples including dairy cows, whereas our study identified O157:H7 in cattle but not pigs [Reference Roug41]. A 2002 study that also collected faecal samples at fairs in three states, identified an E. coli O157:H7 prevalence of 11.4% in cattle, 1.2% in swine and 3.6% in sheep and goats, whereas we measured a 5.31% (6/113) E. coli O157:H7 in all cattle (i.e. combined dairy and beef cattle samples) [Reference Keen42]. Animals shown at fairs are subjected to stress and comingling and may experience increased STEC shedding. Differing STEC prevalence in these aforementioned studies may reflect different management practices or other climate or animal-level factors, as reflected in our results. Additionally, since ruminants are the main reservoirs for STEC, our results indicating that STEC prevalence in swine is lower comparatively than the sampled ruminants are in agreement with previous research. However, pigs are still a livestock species of public health concern, as they harbour E. coli O157:H7 as indicated by many studies [Reference Tseng and Fratamico36, Reference Keen42,Reference Ercoli43]. Our model results also showed that cattle (OR 4.61) and sheep (OR 5.29) are significant factors in STEC presence. However, differences in location, laboratory methods and sampling methods make comparison between studies challenging.

More than half of the typed isolates in this study belonged to an O-group that is listed on the CDC list of the top 7 STEC of concern for public health, including six O157:H7, 22 O26, nine O103 and one O111 [23]. Stx2 was identified in 16.46% (13/79) of isolates, and 13.92% (11/79) contained both stx1 and stx2. Subtyping of stx1 and stx2 was not conducted in this study, therefore direct implications to public health due to virulence subtypes cannot be inferred. The eaeA gene, which allows STEC bacteria to attach to human host cells, was detected in 55.69% (44/79) of positive STEC samples, contrary to a study conducted by Dewsbury et al. (2015), which rarely reported eaeA in their non-O157 isolates from cattle faecal samples [Reference Gonzalez44–Reference Dewsbury47]. The ehxA gene, which is reported in severe human cases of STEC infection, was detected in 88.61% of the positive isolates (70/79) [Reference Djordjevic46]. Compared to a study conducted by Djordjevic et al. (2004) in adult sheep and lambs, they detected stx1, stx2 and ehxA in 78.2% of their positive isolates, vs. our study which only identified those three genes in 1.27% (1/79) of positive isolates. However, they reported 0.8% of their isolates had just stx2 and ehxA genes, whereas, in this current study, 11.39% (9/79) of the positive isolates contained these two virulence genes [Reference Djordjevic46]. The pathogenic STEC O-serogroups, genes and virulence factors identified in this study highlight the need for continued studies on DSSF, as well as the implementation of outreach efforts to stakeholders regarding pre-harvest food safety risks and the development of on-farm mitigation strategies.

Significant risk factors identified by the final mixed-effect model included daily maximum temperature °C, which ranged from 11.7 to 39.80 °C during the study period. An experiment that measured the decline of E. coli O157:H7 in inoculated manure at four temperatures, 7, 16, 23 and 33 °C, reported that E. coli O157:H7 declined significantly faster in manure at 23 and 33 °C than at 7 and 16 °C, for both oscillating and constant temperatures [Reference Semenov48]. These study results confirm our model findings, which suggested that as the daily maximum temperature increased within our temperature range, the odds of finding STEC in a faecal sample were less likely. A study by Franklin et al. (2013) also identified daily maximum temperature as a significant risk factor, when conducting a study of STEC in wild ungulates in Colorado [Reference Franklin49]. They detected a positive association between temperature and STEC presence in faecal samples, whereas our study identified a negative association with the daily maximum temperature [Reference Franklin49]. However, the range of daily maximum temperatures in their analysis was narrower than our recorded temperatures, which may account for this difference. Although many studies indicated that higher STEC shedding occurs during summer months, California microclimates differ from each other and from the majority of seasons in other states [32]. California valleys and foothills experience low humidity and temperatures above 37.78 °C in the summer and autumn, which may affect STEC shedding [39]. For example, to compensate for the numerous microclimates in California in our study on Campylobacter spp., which included the same farms from this current study, we divided the California summer season into Coastal and Inland and the season was a significant risk factor in that final multilevel logistic regression model [Reference Pires11].

Differences in weather (i.e. humidity, temperature range) between states in the USA reveal a need to report the full range of measurements collected for studies estimating the effect of seasons on STEC shedding in livestock. For instance, a study that collected samples from conventional dairy and beef cattle in Michigan revealed that high average temperatures (more than 28.9 °C) measured 1–5 days before sampling had a 2.5 times greater odds of STEC than lower temperatures, which differs from our study results that suggested that STEC survival is less likely at higher maximum temperatures [Reference Venegas-Vargas50]. These Michigan results contradicted ours; however, the highest maximum daily temperature measured in our study (i.e. 39.8 °C) is not a temperature normally observed in many areas of the USA. The range of daily maximum temperatures for the Michigan study was 22.78–32.2 °C, with one 36.11 °C outlier. Additionally, our study included winter temperatures, while their study was only conducted in summer, and summers in Michigan include precipitation events, while little rainfall is observed during California summers. High temperatures or other weather discrepancies observed in different parts of the world may affect conclusions and interpretations of results.

Stanford et al. (2017) reported the effects of severe weather events on STEC shedding in Canadian cattle [Reference Stanford51]. Although they also observed that STEC prevalence increased when ambient temperatures were higher than 28.9 °C, a separate finding indicated that the O-serogroup O26 had a significant reduction in prevalence during extreme heat in July and August [Reference Stanford51]. Almost 28% of the O-serogroups in our study were O26, and the final model results may have been influenced by this strain. The survival of non-O157 STEC strains under different environmental conditions may account for variations in results between studies [Reference Stanford51]. Moreover, physiological and behavioural changes in the host species during various temperature fluctuations or extreme weather events should also be considered [Reference Dawson52]. For instance, Dawson et al. (2018) measured behavioural changes in cattle during increased temperatures, such as increased water consumption or change in grazing habits, as a possible driver of varying STEC prevalence [Reference Dawson52]. Their simulation results indicated that higher summer temperatures may encourage more resting by cattle in crowded areas, such as under shade trees, which can lead to direct transmission of STEC. Since the aforementioned studies differ in conclusions regarding the direction of temperature on STEC shedding in livestock, this risk factor needs further investigation as perhaps there are underlying mechanisms accounting for the difference between results, including microclimates or animal-level factors [Reference Moriarty19, Reference Stanford51].

Our study also indicated that livestock sharing a barn with other animals resulted in 3.5 greater odds of a positive STEC sample. Multiple livestock housed in a barn could share pathogens by cross-contamination of food or water troughs or persistence of STEC in a barn environment that may not be subjected to regular cleaning [Reference Doyle and Erickson17, Reference Gunn53,Reference Rapp54]. Other studies have indicated that STEC persists for long periods of time in barns or on surfaces within a farm environment. For instance, one study swabbed multiple barn surfaces at a dairy ranch and measured 14.9–19.1% STEC in samples from feeders and 11.3–18.0% on other surfaces [Reference Rahn55]. Another study implicated water troughs as harbouring E. coli O157:H7 and inferred that shared water troughs play a role in the persistence and maintenance of continued E. coli O157:H7 infections in cattle [Reference LeJeune, Besser and Hancock56]. A British study reported that housed beef cattle shed more STEC than unsheltered and suggested that this may be due to shared water sources or feeding bins or an accumulation of pathogens in a shared environment [Reference Gunn53].

Finally, the final multivariable model included livestock in contact with wild areas, such as forests or wetlands, as a risk factor for the presence of STEC in faecal samples. Wildlife including feral pigs, deer and birds are known reservoirs of STEC [Reference Kim, Lee and Kim57–Reference Navarro-Gonzalez61]. A study conducted in California identified a low prevalence of E. coli O157:H7 in rodents (0.2%); however, they did not test for non-O157 STEC in samples, which may have a higher prevalence in rodents [Reference Kilonzo62]. A 2016 published study discovered the stx2 gene in over 20% of Canada geese faecal samples and 7% of nearby water samples from Lake Eric bordering Ohio, USA [Reference Hsu58]. Another study of wild birds sampled on a California produce farm detected a low prevalence of 0.34% for O157:H7 and 0.5% for non-O157 STEC, which included O-serogroups O103 and O26 [Reference Navarro-Gonzalez61]. A case-control study conducted after 15 human cases of E. coli O157:H7 identified the source of STEC as those who ate fresh strawberries contaminated by deer faeces [Reference Laidler60]. Livestock that graze in wild areas may be exposed to indirect sources of STEC, for instance through environmental contamination of soil or water, or because wildlife that live in these bordering wild areas enter agricultural areas and contaminate the pastures grazed by animals [Reference Strawn24, Reference Byappanahalli63].

Limitations of this study included a small sample size of farms and an inability to randomise sampled farms, so the model results are not generalisable to other regions and farms. Variables that should be included in future studies to increase our knowledge regarding carriage of foodborne pathogens on DSSF include the age of the animal and whether livestock have direct or indirect contact with neighbouring livestock.

Conclusion

Many consumers perceive DSSF and outdoor-raised livestock as safer than food grown on large-scale conventional farms or meat animals raised in confinement systems. However, identification of STEC O-serogroups that are of public health concern indicates the need for mitigation strategies on DSSF, such as housing livestock species separately and restricting access to wild areas, in order to prevent the transmission of foodborne pathogens in a pre-harvest environment.

Acknowledgements

The authors thank all the farmers and ranchers who confidentially participated in this study. We also thank Dr Irina Udaltsova for statistical advice, and the staff at Western Center for Food Safety and Pires lab for fieldwork and technical support.

Financial support

This project was supported by start-up funds provided by the School of Veterinary Medicine, the Division of Agriculture Natural Resources (UC ANR), University of California, and by the National Institute of Food and Agriculture, US Department of Agriculture, under Award No. 2011-51181-30767. Any opinions, findings, conclusions or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the US Department of Agriculture.

Conflict of interest

None.

Data availability statement

The data that support the findings of this study are available from the corresponding author on reasonable request.

Footnotes

*

Current Affiliation: INRAE, Oniris, BIOEPAR, 44300, Nantes, France.

References

Hilimire, K (2011) Integrated crop/livestock agriculture in the United States: a review. Journal of Sustainable Agriculture 35, 376–393.Google Scholar
Martinez, S et al. (2010) Local food systems: concepts, impacts and issues; USDA-ERS(97).Google Scholar
United States Department of Agriculture, National Agriculture Statistics Service (2015) Certified organic survey 2015 summary. Available at https://www.nass.usda.gov/ (Accessed 22 September 2017).Google Scholar
United States Department of Agriculture, Economic Research Service (2013) The revised ERS farm typology: classifying U.S. farms to reflect today's agriculture. Available at https://www.ers.usda.gov/amber-waves/2013/may/the-revised-ers-farm-typology-classifying-us-farms-to-reflect-todays-agriculture/ (Accessed 22 September 2017).Google Scholar
Kremen, C, Iles, A and Bacon, C (2012) Diversified farming systems: an agroecological, systems-based alternative to modern industrial agriculture. Ecology and Society 17, 44.CrossRefGoogle Scholar
Cesca, J et al. (2016) California Department of Food and Agriculture, California agricultural statistics review, 2015–2016. Available at https://www.cdfa.ca.gov/statistics/PDFs/2016Report.pdf (Accessed 22 September 2017).Google Scholar
Salaheen, S et al. (2015) Zoonotic bacterial pathogens and mixed crop-livestock farming. Poultry Science 94, 1398–1410.CrossRefGoogle ScholarPubMed
Hoar, BR et al. (2013) Buffers between grazing sheep and leafy crops augment food safety. California Agriculture 67, 104–109.CrossRefGoogle Scholar
United Nations. The 203 agenda for sustainable development: Goal 12 Department of Economic and Social Affairs. Available at https://sdgs.un.org/goals/goal12 (Accessed 7 November 2021).Google Scholar
Harvey, RR, Zakhour, CM and Gould, LH (2016) Foodborne disease outbreaks associated with organic foods in the United States. Journal of Food Protection 79, 1953–1958.CrossRefGoogle ScholarPubMed
Pires, AFA et al. (2019) Prevalence and risk factors associated with Campylobacter spp. and Salmonella enterica in livestock raised on diversified small-scale farms in California. Epidemiology and Infection 147, 1–9.CrossRefGoogle ScholarPubMed
Filippitzi, ME et al. (2018) Review of transmission routes of 24 infectious diseases preventable by biosecurity measures and comparison of the implementation of these measures in pig herds in six European countries. Transboundary Emerging Diseases 65, 381–398.CrossRefGoogle ScholarPubMed
Doyle, MP and Erickson, MC (2012) Opportunities for mitigating pathogen contamination during on-farm food production. International Journal of Food Microbiology 152, 54–74.CrossRefGoogle ScholarPubMed
Paletta, ACC, Castro, VS and Conte-Junior, CA (2020) Shiga toxin-producing and enteroaggregative Escherichia coli in animals, foods, and humans: pathogenicity mechanisms, detection methods, and epidemiology. Current Microbiology 77, 612–620.CrossRefGoogle ScholarPubMed
Ongeng, D et al. (2014) Modeling the fate of Escherichia coli O157:H7 and Salmonella enterica in the agricultural environment: current perspective. Journal of Food Science 79, R421–R427.Google ScholarPubMed
Jacob, ME et al. (2013) Evidence of non-O157 Shiga toxin-producing Escherichia coli in the feces of meat goats at a U.S. slaughter plant. Journal of Food Protection 76, 1626–1629.Google Scholar
Doyle, MP and Erickson, MC (2006) Reducing the carriage of foodborne pathogens in livestock and poultry. Poultry Science 85, 960–973.CrossRefGoogle ScholarPubMed
Fraser, EM et al. (2013) Effects of seasonality and a diet of brassicas on the shedding of Escherichia coli O157 in sheep. Foodborne Pathogens and Disease 10, 649–654.CrossRefGoogle Scholar
Moriarty, EM et al. (2011) Survival of Escherichia coli, Enterococci, and Campylobacter spp. in sheep feces on pastures. Applied and Environmental Microbiology 77, 1797–1803.CrossRefGoogle ScholarPubMed
Centers for Disease Control and Prevention (2014) Surveillance for foodborne disease outbreaks, United States, 2012. Annual Report. Atlanta, Georgia: US Department of Health and Human Services, CDC. Available at https://www.cdc.gov/foodsafety/pdfs/foodborne-disease-outbreaks-annual-report-2012-508c.pdf (Accessed 22 September 2017).Google Scholar
Conrad, C et al. (2016) Shiga toxin-producing Escherichia coli and current trends in diagnostics. Animal Frontiers 6, 37–43.CrossRefGoogle Scholar
Mughini-Gras, L et al. (2018) Attribution of human infections with Shiga toxin-producing Escherichia coli (STEC) to livestock sources and identification of source-specific risk factors, The Netherlands (2010–2014). Zoonoses and Public Health 65, e8–22.CrossRefGoogle Scholar
Centers for Disease Control and Prevention (2018) National Shiga toxin-producing Escherichia coli (STEC) surveillance annual report, 2016. Available at https://www.cdc.gov/ecoli/surv2016/index.html (Accessed 20 August 2020).Google Scholar
Strawn, LK et al. (2013) Landscape and meteorological factors affecting prevalence of three food-borne pathogens in fruit and vegetable farms. Applied and Environmental Microbiology 79, 588–600.CrossRefGoogle ScholarPubMed
Food Drug and Administration (2020) Factors potentially contributing to the contamination of romaine lettuce implicated in the three outbreaks of E. coli O157:H7 during the fall of 2019. Available at https://www.fda.gov/food/outbreaks-foodborne-illness/factors-potentially-contributing-contamination-romaine-lettuce-implicated-three-outbreaks-e-coli (Accessed 20 August 2021).Google Scholar
United States Department of Agriculture Forest Service. CalVeg Mapping Zones. Available at https://www.fs.usda.gov/detail/r5/landmanagement/resourcemanagement/?cid=stelprdb5347192 (Accessed 22 September 2017).Google Scholar
Ausvet. Epitools – epidemiological calculators. Available at https://epitools.ausvet.com.au (Accessed 22 September 2017).Google Scholar
Cooley, MB et al. (2013) Development of a robust method for isolation of Shiga toxin-positive Escherichia coli (STEC) from fecal, plant, soil and water samples from a leafy greens production region in California. PLoS ONE 8, e65716.CrossRefGoogle ScholarPubMed
Philpott, D and Ebel, F (2003) E. coli: Shiga toxin methods and protocols biomed protocols. Methods in Molecular Medicine 73, 45–52. (Publisher Springer Science & Business Media).Google Scholar
DebRoy, C, Roberts, E and Fratamico, PM (2011) Detection of O antigens in Escherichia coli. Animal Health Research Reviews 12, 169–185.CrossRefGoogle Scholar
California Department of Water Resources. California Irrigation Management Information System. Available at https://cimis.water.ca.gov/ (Accessed 22 September 2017).Google Scholar
United States Department of Agriculture (2012) USDA Plant Hardiness Zone Map. Available at https://planthardiness.ars.usda.gov (Accessed 22 September 2017).Google Scholar
Douglas, B, Martin, M and Ben Bolker, SW (2015) Fitting linear mixed-effects models using lme4. Journal of Statistical Software 67, 1–48.Google Scholar
Hartig, F and Lohse, L (2021) Package ‘DHARMa’: residual diagnostics for hierarchical (multilevel / mixed) regression models. Available at https://cran.r-project.org/web/packages/DHARMa/DHARMa.pdf (Accessed 20 August 2021).Google Scholar
RStudio (2019) Integrated development for R. RStudio Boston, MA: RStudio, PBC; 2019. Available at http://www.rstudio.com/ (Accessed 10 August 2020).Google Scholar
Tseng, M and Fratamico, PM (2015) Shiga toxin-producing Eschericia coli in swine: the public health perspective. Animal Health Research Review 15, 63–75.CrossRefGoogle Scholar
Cooley, M et al. (2007) Incidence and tracking of Escherichia coli O157:H7 in a major produce production region in California. PLoS ONE 2, e1159.CrossRefGoogle Scholar
Tseng, M et al. (2015) Shiga toxin-producing E. coli (STEC) in swine: prevalence over the finishing period and characteristics of the STEC isolates. Epidemiology and Infection 143, 505–514.CrossRefGoogle ScholarPubMed
Unided States Department of Agriculture APHIS Veterinary Services (2003) Escherichia coli O157 on US Dairy Operations. Available at https://www.aphis.usda.gov/animal_health/nahms/dairy/downloads/dairy02/Dairy02_is_Ecoli_1.pdf (Accessed 22 September 22, 2017).Google Scholar
Patterson, L et al. (2018) Persistence of Escherichia coli in the soil of an organic mixed crop-livestock farm that integrates sheep grazing within vegetable fields. Zoonoses and Public Health 65, 887–896.CrossRefGoogle ScholarPubMed
Roug, A et al. (2013) Zoonotic fecal pathogens and antimicrobial resistance in county fair animals. Comparative Immunology Microbiology and Infectious Diseases 36, 303–308.CrossRefGoogle ScholarPubMed
Keen, JE et al. (2006) Shiga-toxigenic Escherichia coli O157 in agricultural fair livestock, United States. Emerging Infectious Diseases 12, 780–786.CrossRefGoogle ScholarPubMed
Ercoli, L et al. (2015) Role of verocytotoxigenic Escherichia coli in the swine production chain. Italian Journal of Food Safety 4, 116–122.CrossRefGoogle ScholarPubMed
Gonzalez, AGM et al. (2016) Serotypes, virulence markers and cell invasion ability of Shiga toxin-producing Escherichia coli strains isolated from healthy dairy cattle. Journal of Applied Microbiology 121, 1130–1143.CrossRefGoogle ScholarPubMed
Beutin, L et al. (2004) Characterization of Shiga toxin-producing Escherichia coli strains isolated from human aptients in Germany over a 3-year period. Journal of Clinical Microbiology 42, 1099–1108.CrossRefGoogle Scholar
Djordjevic, SP et al. (2004) Serotypes and virulence gene profiles of Shiga toxin-producing Escherichia coli strains isolated from feces of pasture-fed and lot-fed sheep. Applied and Environmental Microbiology 70, 3910–3917.CrossRefGoogle ScholarPubMed
Dewsbury, DMA et al. (2015) Summer and winter prevalence of Shiga toxin-producing Escherichia coli (STEC) O26, O45, O103, O111, O121, O145, and O157 in feces of feedlot cattle. Foodborne Pathogens and Disease 12, 726–732.CrossRefGoogle Scholar
Semenov, AV et al. (2007) Influence of temperature fluctuations on Escherichia coli O157:H7 and Salmonella enterica serovar Typhimurium in cow manure. FEMS Microbiology Ecology 60, 419–428.CrossRefGoogle ScholarPubMed
Franklin, AB et al. (2013) Wild ungulates as disseminators of Shiga toxin-producing Escherichia coli in urban areas. PLoS ONE 8, 8–13.CrossRefGoogle ScholarPubMed
Venegas-Vargas, C et al. (2016) Factors associated with Shiga toxin-producing Escherichia coli shedding by dairy and beef cattle. Applied and Environmental Microbiology 82, 5049–5056.CrossRefGoogle ScholarPubMed
Stanford, K et al. (2017) Effect of severe weather events on the shedding of Shiga toxigenic Escherichia coli in slaughter cattle and phenotype of serogroup O157 isolates. FEMS Microbiology Ecology 93, 1–12.Google ScholarPubMed
Dawson, DE et al. (2018) Investigating behavioral drivers of seasonal Shiga-toxigenic Escherichia coli (STEC) patterns in grazing cattle using an agent-based model. PLoS ONE 13, 1–24.CrossRefGoogle ScholarPubMed
Gunn, GJ et al. (2007) An investigation of factors associated with the prevalence of verocytotoxin producing Escherichia coli O157 shedding in Scottish beef cattle. The Veterinary Journal 174, 554–564.CrossRefGoogle ScholarPubMed
Rapp, D et al. (2021) Investigation of on-farm transmission routes for contamination of dairy cows with top 7 Escherichia coli O-serogroups. Microbial Ecology 81, 67–77.CrossRefGoogle ScholarPubMed
Rahn, K et al. (1997) Persistence of Escherichia coli O157:H7 in dairy cattle and the dairy farm environment. Epidemiology and Infection 119, 251–259.CrossRefGoogle ScholarPubMed
LeJeune, JT, Besser, TE and Hancock, DD (2001) Cattle water troughs as reservoirs of Escherichia coli O157. Applied and Environmental Microbiology 67, 3053–3057.CrossRefGoogle ScholarPubMed
Kim, JS, Lee, MS and Kim, JH (2020) Recent updates on outbreaks of Shiga toxin-producing Escherichia coli and its potential reservoirs. Frontiers in Cellular and Infection Microbiology 10, 1–10.CrossRefGoogle ScholarPubMed
Hsu, TTD et al. (2016) Prevalence and diversity of Shiga-toxin genes in Canada geese and water in western Lake Erie Region. Journal of Great Lakes Research 42, 476–481.CrossRefGoogle Scholar
Brown, VR, Bowen, RA and Bosco-lauth, AM (2018) Zoonotic pathogens from feral swine that pose a significant threat to public health. Transboundary and Emerging Diseases 65, 649–659.CrossRefGoogle Scholar
Laidler, MR et al. (2013) Escherichia coli O157:H7 infections associated with consumption of locally grown strawberries contaminated by deer. Clinical Infectious Diseases 57, 1129–1134.CrossRefGoogle ScholarPubMed
Navarro-Gonzalez, N et al. (2020) Carriage and subtypes of foodborne pathogens identified in wild birds residing near agricultural lands in California: a repeated cross-sectional study. Applied and Environmental Microbiology 86, e01678–19.CrossRefGoogle ScholarPubMed
Kilonzo, C et al. (2013) Fecal shedding of zoonotic food-borne pathogens by wild rodents in a major agricultural region of the Central California Coast. Applied and Environmental Microbiology 79, 6337–6344.CrossRefGoogle Scholar
Byappanahalli, MN et al. (2006) Population structure, persistence, and seasonality of autochthonous Escherichia coli in temperate, coastal forest soil from a Great Lakes watershed. Environmental Microbiology 8, 504–513.CrossRefGoogle ScholarPubMed
Figure 0

Table 1.Virulence genes discovered in 79 isolates from 76 positive Shiga toxin-producing Escherichia coli samplesa collected from 16 California diversified small-scale farms, between May 2015 and June 2016, characterised by livestock species, O-serogroup and virulence gene

Figure 1

Table 2.Mean, median and range (i.e. minimum and maximum) of selected continuous variables assessed for model building and collected during a cross-sectional study conducted from 2015 to 2016 on 14a diversified small-scale farms in California

Figure 2

Table 3.Bivariate analysis for selected categorical variables stratified by positive (n = 69) or negative (n = 433) Shiga toxin-producing Escherichia coli status samples were collected during a cross-sectional study from 2015 to 2016 on 14c diversified small-scale farms in California

Figure 3

Table 4.Association between risk factors and the presence of Shiga toxin-producing Escherichia coli in faecal samples (n = 502) collected from 14a diversified small-scale farms in California between May 2015 and June 2016, as demonstrated by a generalised linear mixed model, with farm as a random effect