P3-217 Comparison of Two Methods for Enumeration of Total Fecal Coliforms and Generic Escherichia coli, and Their Ability to Predict Pathogen Occurrence in Irrigation Waters

Wednesday, July 12, 2017
Exhibit Hall (Tampa Convention Center)
Justin Falardeau , Food, Nutrition and Health, University of British Columbia , Vancouver , Canada
Roger Johnson , National Microbiology Laboratory at Guelph , Guelph , Canada
Siyun Wang , Food, Nutrition and Health, University of British Columbia , Vancouver , Canada
Introduction: Current methods for assessing quality of irrigation water in Canada and the United States are based on the hygiene indicators, fecal coliforms (FC) and generic Escherichia coli (EC); however, debate exists as to whether these indicators are predictive of pathogen occurrence.

Purpose: Our objective was to compare the results from two different FC/EC enumeration methods and determine their ability to predict the presence of verotoxigenic E. coli (VTEC), Salmonella, and Listeria monocytogenes in irrigation water.

Methods: Water was collected bi-monthly over a 15-month period from seven irrigation ditches within two distinct watersheds in British Columbia, Canada. FC and EC were enumerated by 1) a 1 ml aliquot on 3M™ Petrifilm™ E. coli/Coliform count plates, and 2) a 25 ml membrane filtration with growth on m-FC agar with 0.01% rosalic acid, followed by transfer to nutrient agar containing 4-methylumbelliferyl-β-D-glucuronide.  The samples were tested for the presence of VTEC using a verotoxin colony immunoblot (VT-IB), and for Salmonella and L. monocytogenes using Health Canada methods.

Results: The two methods were found to be significantly correlated for observed counts of both FC (r = 0.60; p < 0.001) and EC (r = 0.77; p < 0.001); however, the Petrifilm™ method gave higher numbers in general for both indicators. Enumeration by membrane filtration significantly correlated with pathogen presence for both FC (r = 0.431; p < 0.001) and EC (r = 0.408; p < 0.001), but the Petrifilm™ method only correlated significantly for FC (r = 0.234; p < 0.001).  Regression tree analysis showed a split of an 83.3% chance of pathogen occurrence in samples with ≥ 7.54 CFU/ml FC when measured by membrane filtration.

Significance: This research shows that while FC and EC may correlate with the presence of pathogens in water, the enumeration method and volume tested may play a role in their accuracy.