Evaluating the U.S. Food Safety Modernization Act Produce Safety Rule Standard for Microbial Quality of Agricultural Water for Growing Produce

Abstract

ABSTRACT The U.S. Food and Drug Administration (FDA) has defined standards for the microbial quality of agricultural surface water used for irrigation. According to the FDA produce safety rule (PSR), a microbial water quality profile requires analysis of a minimum of 20 samples for Escherichia coli over 2 to 4 years. The geometric mean (GM) level of E. coli should not exceed 126 CFU/100 mL, and the statistical threshold value (STV) should not exceed 410 CFU/100 mL. The water quality profile should be updated by analysis of a minimum of five samples per year. We used an extensive set of data on levels of E. coli and other fecal indicator organisms, the presence or absence of Salmonella, and physicochemical parameters in six agricultural irrigation ponds in West Central Florida to evaluate the empirical and theoretical basis of this PSR. We found highly variable log-transformed E. coli levels, with standard deviations exceeding those assumed in the PSR by up to threefold. Lognormal distributions provided an acceptable fit to the data in most cases but may underestimate extreme levels. Replacing censored data with the detection limit of the microbial tests underestimated the true variability, leading to biased estimates of GM and STV. Maximum likelihood estimation using truncated lognormal distributions is recommended. Twenty samples are not sufficient to characterize the bacteriological quality of irrigation ponds, and a rolling data set of five samples per year used to update GM and STV values results in highly uncertain results and delays in detecting a shift in water quality. In these ponds, E. coli was an adequate predictor of the presence of Salmonella in 150-mL samples, and turbidity was a second significant variable. The variability in levels of E. coli in agricultural water was higher than that anticipated when the PSR was finalized, and more detailed information based on mechanistic modeling is necessary to develop targeted risk management strategies.