UC Davis

Background. California Leafy Green Products Handler Marketing Agreement (LGMA) established food safety metrics for producing leafy greens, with guidance recommendations of 400 feet, 1200 feet, and one-mile distances between production fields and either a composting facility utilizing animal products, or a feedlot (or concentrated animal feeding operations (CAFO)) containing >1000 or >80,000 head of cattle, respectively. Aim. The purpose of the three chapters was to: (1) evaluate the effect of these distance metrics, (2) identify key environmental risk factors associated with airborne bacterial pathogens (E. coli O157, non-O157 Shiga-toxin–producing E. coli (STEC), Salmonella) and indicator E. coli in proximity to beef cattle CAFOs, and (3) compare the cultural methods of airborne E. coli and the sequence of uspA gene from airborne E. coli isolated in each feedlot. Methods. For chapter 1, each sample contained 1000 liters of air at a 1.2-m elevation over 10 minutes using MAS-100 Eco microbial air samplers. Airborne E. coli was tested based on direct count. Meteorological data in situ (air temperature, wind speed, wind direction, and relative humidity) was collected. Logistic regression was used to identify the association between risk factors and the odds of airborne E. coli detection. For chapter 2, each sample comprised 6000 liters of air collected at 1.2-m elevation using the same air samplers, with TSB-enriched air filters qPCR-screened for E. coli O157, STEC, Salmonella, and indicator E. coli; suspect positive colonies were further qPCR-confirmed. A separate air sample was collected for direct enumeration of the concentration of indicator E. coli. Local meteorological data was collected in situ and from a nearby weather station, along with the line-of-sight distance from the feedlot and events of dust-generating activity, with logistic regression used to identify which of these factors were associated with the odds of bacterial detection. For chapter 3, McNemar’s test was applied to compare the two methods (direct count and TSB-enrichment) of detecting airborne E. coli. The sequence similarity of the uspA gene from airborne E. coli tested positive based on both direct count and TSB-enrichment was calculated based on neighbors-joining methods. Wilcoxon rank-sum test was used for the association between the land occupation of feedlot and the occurrence of airborne E. coli. The sequence of the uspA gene from airborne E. coli in each feedlot was analyzed based on phylogenetic analysis. Results. For chapter 1, a total of 168 air samples were collected from seven beef cattle feedlots in March and April 2020. The distance away from the edge of the feedlot ranged from 0 to around 2200 m (1.3 miles). The prevalence of airborne E. coli is 6.5% (11/168). The concentration of the airborne E. coli ranged from 1 to 2 colony-forming unit (CFU) per 1000 L of air. All positive samples are 37 m (120 ft) from the edge of the feedlot. As the distance between the feedlot edge and the air sampler increases 100 ft (30 m), the odds of detecting E. coli decrease by 95% (OR=0.05). For chapter 2, during the 6-month period (11/2020 to 4/2021) for leafy green production in Imperial Valley, California, 150 air samples were collected at LGMA-guidance distances of ~400 feet, ~1200 feet, and ~1 mile from five cattle feedlots. In addition, 150 samples were collected at randomized distances from the five feedlots ranging from 30 to 2000 feet in all directions. No bacterial pathogens were qPCR-confirmed for the 300 samples totaling 1.8 million L of processed air, which suggests a maximum concentration for E. coli O157, non-O157 STEC, and Salmonella of less than ~1 CFU per million L of ambient air. Indicator E. coli was detected in 16.7% (50/300) of samples, with positives found at all distance categories and a concentration ranging from 0 to 19 CFU/6000 L. Although E. coli detection was not significantly different between air samples taken at LGMA-guidance distances of 400 ft, 1200 ft, and 1 mile (8.3%, 15.0%, 10.0%, respectively), logistic regression on all 300 air samples (n=150 random plus n=150 LGMA-fixed distances) found significantly higher odds of E. coli for samples taken in close proximity compared to >2000-ft distance from a feedlot. Environmental factors associated with E. coli detection included wind speed, wind direction, relative humidity, sampling hour of day and month, and the presence of activities that created fugitive dust (feedlot cattle activity, vehicular traffic, plowing fields) during the sampling period. For chapter 3, statistically, there is a significant difference between the two test methods (direct count and TSB-enrichment) for airborne E. coli (p<0.001). For the 12 samples, which were positive for airborne E. coli based on both direct count and enrichment cultural methods, the sequence similarity of the uspA gene ranged from 88.6% to 100% with a mean of 93.1%. With comparable data, the prevalence of airborne E. coli for the first and second studies is 1.99% and 7.69%, respectively. Statistically, the occurrence of airborne E. coli is greater in the second study compared to the first one (p=0.04). The difference is from the volume of air collected, the duration of sampling, and the air filter placed on top of the media. Land occupancy of feedlots (an ordinal variable) is not associated with the occurrence of airborne E. coli for the first (p=0.07) and the second study (p=0.28) within 1 mile of feedlots. The sequence similarity of the uspA gene from all the airborne E. coli isolates ranged from 88.6% to 100% in each feedlot. If treating E. coli isolated from ≤50 ft and intensively exposed to a feedlot as reference isolates, 1400 ft is the furthest distance for E. coli isolates that shared identical uspA gene with reference isolates. For all the airborne E. coli isolated from >50 ft from the edge of each feedlot, ~6% to ~60% of the isolates shared identical sequence of the uspA gene with the reference isolates. Conclusion. Lack of bacterial pathogen detection within these sampled distances suggests airborne deposition from nearby feedlots may not be a significant source of leafy green bacterial pathogen contamination; detection of very low concentrations of indicator E. coli as a function of distance, wind speed, and direction provides data to inform future revisions of produce safety guidance documents such as the current version of the California LGMA.