Modeling Salmonella Growth from a Low Initial Density on Chicken Products with Native Microflora

1
Modeling Salmonella Growth from a Low Initial
Density on Chicken Products with Native
Microflora Thomas P. Oscar, Agricultural Research
Service, USDA, Room 2111, Center for Food Science
and Technology, University of Maryland Eastern
Shore, Princess Anne, MD 21853 410-651-6062
410-651-8498 (fax) thomas.oscar_at_ars.usda.gov
RESULTS AND DISCUSSION Variation of Salmonella
growth among portions and batches of chicken skin
with native microflora was significant (Fig. 1).
This necessitated the development of a stochastic
model.
INTRODUCTION Lack of Salmonella strains with
phenotypes that can be quantified in the presence
of other microorganisms and lack of rapid methods
for quantifying low initial levels of Salmonella
are two technical issues that prevent development
of predictive microbiology models in chicken
products with native micro-flora and from a low
and ecological dose of Salmonella. Although
transformation of Salmonella with a green
fluorescent protein gene from a jelly fish
produces strains with a phenotype that can be
quantified in the presence of other
microorganisms, the transformed strains grow
slower than the parent strains. Moreover, the
slower growth is affected by a strain by
temperature by growth parameter interaction that
prevents application of a correction factor that
would allow use of the strains for model
development. Recently, it was discovered that a
multiple antibiotic resistant (MAR) strain of
Salmonella Typhimurium DT104 that occurs in
nature has a phenotype that can be quantified in
chicken products with native micro-flora. This
strain was found to grow similar to other
Salmonella strains from poultry and its MAR
phenotype allowed development of a rapid MPN
assay for development and validation of models
for growth of Salmonella from a low initial dose
on chicken products (i.e. ground chicken breast
meat and chicken frankfurters) with native
micro-flora. OBJECTIVE To develop and validate
a model for growth of Salmonella from a low
initial dose on chicken skin with a native
micro-flora. MATERIALS AND METHODS Organism.
A multiple antibiotic resistant (MAR) strain
(ATCC 700408) of Salmonella Typhimurium DT104 was
used for model development and validation. Chicke
n skin preparation. Non-kosher chicken thighs
were purchased from local retail outlets, whereas
kosher chicken thighs were purchased online.
Chicken thigh skin was frozen briefly and then
cut into circular 2.14 cm2 portions. Chicken
skin inoculation and incubation. Chicken skin
portions were spot inoculated (5 ml) with 0.8
log10 S. Typhimurium DT104 followed by incubation
at 5 to 50oC on top of thigh meat in plastic
jars. At selected times of incubation, an
inoculated portion was added to 9 ml of buffered
peptone water, pulsified and then the number of
Salmonella in the pulsifate was determined using
a 3 x 4 MPN method for low cell counts (0 to 3.24
log) or a spiral plating method for higher cell
counts (gt 3 log). Modeling. Kinetic data were
graphed as a function of time and were fit to the
Baranyi primary model to determine lag time,
growth rate and the 95 prediction interval,
which characterized the uncertainty of the
primary model fit as well as the variation of
Salmonella growth among batches of chicken skin.
Secondary models for primary model parameters as
a function of temperature were developed by
non-linear regression and then combined with the
primary model in a spreadsheet to create a
tertiary model that predicted the variation of
Salmonella growth as a function of time and
temperature. Performance of the tertiary model
was evaluated against both dependent and
independent data using the acceptable prediction
zone method.
The secondary models for 95 prediction interval
(Fig. 2), lag time (Fig. 3) and growth rate (Fig.
4) were combined with the primary model (Fig. 1)
in a computer spreadsheet to create a tertiary
model (Fig. 5) for predicting the growth of
Salmonella Typhimurium DT104 from a low initial
dose on non-kosher chicken skin as a function of
time and temperature.
Fig. 1
Fig. 5
Variation of growth among portions and batches of
chicken skin was quantified using a 95
prediction interval. The 95 prediction interval
increased as a function of temperature and, in
general, was greater under growth than no growth
conditions (Fig. 2).
Performance of the tertiary model was evaluated
against dependent data (Fig. 6) and independent
data for interpolation (Fig. 7) and for
extrapolation to kosher chicken skin (Fig. 8)
using the acceptable prediction zone (APZ)
method. The tertiary model had acceptable
performance (i.e. gt 70 of prediction errors were
in the APZ) for all data sets indicating that it
provided acceptable predictions of Salmonella
Typhimurium DT104 growth on both non-kosher and
kosher chicken skin.
Fig. 2
Fig. 6
APZ 82.6

The minimum temperature for growth of Salmonella
is usually 5.5C or above. In the current study,
the minimum temperature for growth of Salmonella
was 21.5C because the length of the storage trial
was only 8 h (Fig. 3).
Fig. 3
Fig. 7
APZ 83.7
Growth rate of Salmonella Typhimurium DT104 on
chicken skin was optimal at 40C and was found to
fit well to a cardinal temperature model (Fig. 4).
Fig. 8
Fig. 4
APZ 81.6
Current studies are aimed at determining how well
the model can extrapolate to other strains of
Salmonella, other chicken products with native
micro-flora, other initial doses of Salmonella
and other previous histories.
I would like to than Jacquelyn Ludwig, Stacey
Engster and Sharif Walker for their excellent
technical assistance on this project.