Prediction of Most Probable Number of Listeria monocytogenes Using a Generalized Linear Model and a Modified FDA Listeria Isolation Method

Abstract

The U.S. Food and Drug Administration (FDA) Listeria isolation method was modified and used to enumerate low levels of Listeria monocytogenes in a spiked aseptic lowfat milk product by using a three-tube most probable number (MPN) design. A generalized linear model (GNL) (assuming a binomial distribution) was used to predict the MPN of L. monocytogenes in a sample when high populations of either Pseudomonas aeruginosa or Lactobacillus casei were present. The MPN of L. monocytogenes was less than the number of cells spiked into the test product, especially at high cell densities. Lithium chloride phenylethanol moxalactam (LPM) agar and modified Oxford (MOX) agar were equally effective in isolating L. monocytogenes using MPN methodology. Forty-eight hour incubation of the MPN broth increased the sensitivity and accuracy compared to 24-h incubation. The presence of P. aeruginosa did not effect the enumeration of L. monocytogenes, while high numbers of L. casei resulted in low MPN estimates.