A conceptual framework for developing recommendations for no-harvest buffers around in-field feces

Results of previous studies revealed that (i) splash can transfer microbes from in-field feces to preharvest produce and (ii) wildlife can be vectors for the introduction of foodborne pathogens into produce fields. However, few peer-reviewed studies have been conducted to examine pathogen transfer from wildlife feces to in-field produce via splash during irrigation. Although two previous studies found a significant relationship between distance and Escherichia coli transfer via splash, the studies sampled produce < m from the feces. The present study was conducted to refine our understanding of the impact of distance on E. coli splash. Two trials were conducted 1 month apart. For each trial, fecal pellets inoculated with a three-strain E. coli cocktail were placed in a lettuce field 2.5 h before irrigation. After irrigation, E. coli levels on lettuce heads 0 to 6 m from the pellets were determined. Although E. coli was not detected in any of the heads ≥2 m from the fecal pellets (n=39), 39% of heads (13 of 33) < m from the pellets tested positive for E. coli. According to logistic regression, the odds of harvesting a head that tested positive for E. coli decreased by a factor of 50 (odds ratio, 0.02; 95% confidence interval, <,0.01, 0.28; P=0.004) for each meter increase in the distance between the lettuce and the feces. Thus, the likelihood of E. coli transfer from feces to produce should be minimal at a given distance from the feces. Our model can be used to predict the probability of harvesting a microbially contaminated lettuce head following implementation of a no-harvest buffer around in-field feces. For example, our model suggests that the probability of harvesting a contaminated head was 0.1% at 3 m from the feces. Although the approaches utilized in this study provide a conceptual framework that can be used to help define appropriate no-harvest buffers, delineation of appropriate buffer zones requires additional information (e.g., acceptable risk and regional data). HIGHLIGHTS • None of the 39 lettuce heads > 2 m from the feces had detectable levels of E. coli. • Thirteen (39%) of the 33 heads > 2 m from the feces had detectable levels of E. coli. • Odds of harvesting E. coli-contaminated lettuce decrease with distance from feces. • Analyses provide a framework for estimating no-harvest buffers around in-field feces. • Risk assessments are needed to set appropriate buffer zones for different regions.