Recycle pump for studies of kinetics of heterogeneous vapor-phase reactions

Abstract

A precise and accurate method for enumeration of low level of Listeria monocytogenes in foods is critical to a variety of studies. In this study, paired comparison of most probable number (MPN) and direct plating enumeration of L. monocytogenes was conducted on a total of 1730 outbreak-associated ice cream samples that were naturally contaminated with low level of L. monocytogenes. MPN was performed on all 1730 samples. Direct plating was performed on all samples using the RAPID'L.mono (RLM) agar (1600 samples) and agar Listeria Ottaviani and Agosti (ALOA; 130 samples). Probabilistic analysis with Bayesian inference model was used to compare paired direct plating and MPN estimates of L. monocytogenes in ice cream samples because assumptions implicit in ordinary least squares (OLS) linear regression analyses were not met for such a comparison. The probabilistic analysis revealed good agreement between the MPN and direct plating estimates, and this agreement showed that the MPN schemes and direct plating schemes using ALOA or RLM evaluated in the present study were suitable for enumerating low levels of L. monocytogenes in these ice cream samples. The statistical analysis further revealed that OLS linear regression analyses of direct plating and MPN data did introduce bias that incorrectly characterized systematic differences between estimates from the two methods.