Purpose: To compare a novel rinse/membrane filtration method (RMFM) to more traditional sponge rubbing or stomaching in preparing fresh produce samples for detection of pathogenic bacteria.
Methods: Decontaminated jalapeño peppers and cantaloupes were inoculated with three levels of inoculum of each organism separately (S. enterica, L. monocytogenes and E. coli O157:H7; 102, 104 and 106 CFU). Bacteria were recovered using the RMFM in parallel with homogenization by stomacher and a sponge surface rubbing method, followed by standard culture methods (BAM). To determine relative accuracy, relative specificity and relative sensitivity, the MFM was compared to the other methods following the ISO16140:2003 protocol. The detection limit was determined by the MPN technique and results analyzed (Kruskal Wallis and Fisher tests).
Results: When 104 and 106 CFU were inoculated, all bacteria were detected in 100% of samples by the three methods. However, with an inoculum of 102 CFU, S. enterica was detected in 20%, 60% and 20% of samples using RMFM, stomacher and sponge respectively. L. monocytogenes was detected in 40%, 60% and 20% of samples using RMFM, stomacher and sponge respectively; whereas E. coli was detected in 80%, 60% and 60% of samples by RMFM, stomacher and sponge, respectively. The limit of detection of RMFM (1.0 - 2.3 MPN/g) was similar to the other methods (0.9 - 2.8 MPN/g). In all cases, specificity and relative accuracy was more than 95%.
Significance: Overall, the RMFM performed similar or better than the homogenization and sponge surface sampling methods. The RMFM method might be a good alternative for processing large numbers of samples for pathogen detection.