Poult. Sci.
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Poult Sci 2007. 86:191-201
© 2007 Poultry Science Association
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Payne, J. B.
Right arrow Articles by Sheldon, B. W.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Payne, J. B.
Right arrow Articles by Sheldon, B. W.

PRODUCTION, MODELING, AND EDUCATION

Modeling the Growth and Death Kinetics of Salmonella in Poultry Litter as a Function of pH and Water Activity

J. B. Payne*, J. A. Osborne{dagger}, P. K. Jenkins* and B. W. Sheldon*,1

* Department of Poultry Science and {dagger} Department of Statistics, North Carolina State University, Raleigh 27695

1 Corresponding author: brian_sheldon{at}ncsu.edu


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 REFERENCES
 
Contaminated poultry litter, serving as a reservoir for Salmonella, can be linked to both food safety concerns when contaminated birds enter processing plants and environmental concerns when used as a fertilizer. Predictive modeling allows for the estimation of microbial growth or inactivation as a function of controlling environmental growth factors. A study was conducted to observe the combined effects of pH and water activity (Aw) at a constant temperature on Salmonella populations in used turkey litter to predict microbial response over time. Litter, first pH-adjusted and then inoculated with a 3-strain Salmonella serovar cocktail to an initial concentration of ~107 cfu/g, was placed into individual sealed plastic containers with saturated salt solutions for controlling Aw. A balanced design including 3 Aw values (0.84, 0.91, 0.96), 3 pH values (4, 7, 9), and a constant temperature of 30°C was used, with litter samples periodically removed and analyzed for Salmonella populations, pH, and Aw. At each combination of environmental factors, the Churchill or exponential inactivation mathematical models were used to describe the growth and death of Salmonella over time. Salmonella populations exhibited growth (~2 log) with little decline up to 42 d in litter environments of pH 7 and 9 and a Aw of 0.96. As litter Aw and pH levels were reduced, populations declined, with the most drastic reductions (~5 log in 9 h) occurring in low-pH (4) and low-Aw (0.84) environments. Generalized models for bacterial growth and death under grouped pH environments were successfully developed to predict Salmonella behavior in litter over time. These findings suggest that the best management practices and litter treatments that lower litter Aw to ≤0.84 and pH to ≤4 are effective in reducing Salmonella populations. The use of a single equation to predict the growth and decline of Salmonella populations as a function of pH and Aw has potential application for use in the development of effective pathogen control strategies at the farm level.

Key Words: Salmonella • mathematical modeling • litter • pH • water activity


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 REFERENCES
 
Salmonella continues to be a foodborne pathogen of concern and has been linked to the consumption of contaminated poultry products (Tauxe, 1991; Bryan and Doyle, 1995; Hoszowski et al., 1996; Byrd et al., 1997; Cox et al., 2000). Human infection is usually attributed to cross-contamination in the kitchen, inadequate cooking, and improper storage temperatures (Silliker, 1980). Annually, Salmonella alone is responsible for an estimated 300,000 to 4,000,000 cases of foodborne illness in the United States (USDA Food Safety and Inspection Service, 1998) and an estimated 30.6% of the deaths associated with foodborne illness (Kiessling et al., 2002). Not only does Salmonella pose a significant public health threat, but it also has an economic effect from costs associated with disease investigations and testing, health care, loss of productivity, compensation for illness and death (McMeekin et al., 1993), and costs to the poultry industry by threatening consumer markets and increasing production and processing costs (Bender and Mallinson, 1991). Consumer confidence and purchase decisions have a direct correlation to the perceived safety and wholesomeness of the product.

To combat contamination of poultry products, Hazard Analysis Critical Control Point (HACCP) programs have been developed and implemented in all US processing plants under federal inspection as a means to identify and control or eliminate potential food safety hazards. This federally-mandated program calls for the testing of the poultry processing plant environment and carcasses for the presence of Salmonella and generic Escherichia coli. The HACCP rule became effective in large processing plants in 1998, small establishments in 1999, and very small establishments in 2000. Before HACCP implementation, Salmonella contamination in broiler carcasses was estimated at 20%, according to a nationwide broiler baseline survey conducted from 1994 to 1995 by the USDA Food Safety and Inspection Service (FSIS; USDA FSIS, 1995). In 2002, the FSIS reported a Salmonella prevalence of 11.5% on broiler carcasses collected from all size processing establishments, a near 50% reduction since HACCP implementation (USDA FSIS, 2003). Processors must provide control measures for this pathogen, such as applying disinfectants to carcass washers and chillers. However, processors are not held responsible for complete pathogen elimination on uncooked products, because contaminated birds can arrive at plants with either undetectable to heavy pathogen loads. In these latter cases, the use of proper control procedures to reduce carcass pathogen incidence may not always assure compliance with the USDA Salmonella performance standards, as specified in the processor’s HACCP plan and federal regulations.

To successfully meet federal and processing plant pathogen-control standards, recent interest has centered on the implementation of on-farm pathogen reduction programs to reduce contamination loads in and on birds entering the processing plant. Commercially reared birds are in constant contact with litter, which can be a significant reservoir for Salmonella contamination (Bryan et al., 1979; Jones et al., 1991; Bryan and Doyle, 1995; Corrier et al., 1999; Trampel et al., 2000). Previous investigations have documented that the Salmonella serotypes isolated from processed products are generally found in the production house litter and other areas of the production environment (Bains and MacKenzie, 1974; Lahellec and Colin, 1985).

The survival of Salmonella in the poultry house environment is dependent on both physical and chemical factors such as temperature, water activity (Aw) or equilibrium RH (ERH), moisture content, and pH. Whenever extrinsic environmental factors fall outside the optimum range for microbial growth and survival, these factors can cause cellular damage. Depending on the severity of the stress factors, growth can be inhibited or cell death can occur (Farkas, 2001). The food safety hurdle concept is an approach that combines several inhibitory hurdles or stress factors that together can act synergistically to inhibit microbial pathogens (Leistner, 2000). When combined, these hurdles are more effective than when applied individually; therefore, their incorporation is recommended to inhibit microbial growth. The findings from previous studies indicate that extrinsic parameters can influence the presence or absence of Salmonella in broiler litter, with the most significant factor being Aw (Opara et al., 1992). Turnbull and Snoeyenbos (1973) concluded that the salmonellacidal activity of used litter may be attributed to changing litter Aw and pH.

Predictive modeling has been used to estimate an organism’s growth or death in a nutrient broth or defined model food system as a function of specific environmental factors. This approach allows food scientists to predict the effect of altering extrinsic factors or treatments on microbial growth, shelf life, and safety of consumer food products and may help to identify those conditions to avoid. Many growth kinetic models have been described in the literature, such as the Gompertz, log-linear, vitalistic, and Baranyi models. These models may be fit to observed data and used for statistical inference about growth and death rates in microbial populations of interest.

One common disadvantage of these models is that they are not sufficiently complex to accommodate populations exhibiting both growth and decline over time. A model proposed by Churchill and Usagi (1972), called the Churchill model, was developed for use in chemical engineering applications and has been adapted for use in microbial kinetic studies (Membre et al., 1997). This approach allows for both a growth and decline phase in the model.

Although the majority of predictive models have been used in model food systems, there may be opportunity for useful application in other systems, such as live animal production environments. Predictive modeling can serve as a useful tool for assessing Salmonella contamination issues currently facing the poultry industry. By understanding the behavior of Salmonella in the growout environment and how these pathogens are affected by various environmental growth parameters, such as pH, Aw, and temperature, preharvest Salmonella control strategies can be developed for reducing foodborne pathogen prevalence and populations on birds entering processing plants and thus their subsequent transmission to humans.

A study was conducted to observe the combined effects of pH and Aw at a constant temperature on the growth and decline of Salmonella spp. in inoculated poultry litter to predict microbial behavior using statistical modeling. Potential application for developing preharvest production management strategies to control Salmonella is discussed.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 REFERENCES
 
Poultry Litter
Used pine shavings from 6-wk-old turkeys were obtained from the University Lake Wheeler Road Poultry Farm and sterilized by autoclaving at 121°C for 1 h. Sterilization was confirmed by placing 25 g of autoclaved litter into 225 mL of buffered peptone water (BPW; Oxoid Ltd., Ogdensburg, NY) and then spread-plating in duplicate onto brain heart infusion agar (BHI; Difco, Sparks, MD) plates. All plates were incubated at 37°C for 24 h, and the colonies were enumerated. For pH 9 and 4 trials, 1,500 g of litter was pH-adjusted to the desired level, as determined from preliminary studies. A plastic sprayer was used to mist 215 mL of 2 M NaOH for pH 9 and 165 mL of 2 M HCl for pH 4 litter conditions, followed by thorough mixing for 5 min, when inside an autoclave bag (63.5 x 88.9 cm; Fisherbrand Autoclave Bag, no. 01-814-C, Fisher Scientific Int., Pittsburgh, PA). The pH 7 litter trials did not require pH adjustment due to the preexisting neutral pH litter conditions. To assay litter pH, 10-g samples were combined with 100 mL of deionized water, stirred, and allowed to stand for 1 min (Pope and Cherry, 2000). A pH reading (Accumet Model 50 pH/Ion/Conductivity Meter, Denver Instrument Co., Arvada, CO; Accuflow Flushable Junction pH Combination Electrode, no. 13620116, Fisher Scientific Int.) was then recorded. The litter was then placed into a drying oven overnight at 100°C to reach a homogeneous litter Aw before inoculation and Aw adjustment.

Culture Preparation
A Salmonella cocktail composed of Salmonella Heidelberg (ATCC no. 8326, American Type Culture Collection, Manassas, VA), Salmonella Newport (ATCC no. 6962, American Type Culture Collection), and Salmonella Typhimurium (ATCC no. 14028, American Type Culture Collection) was prepared for use as a litter inoculum. An aliquot of each strain was rehydrated with 10 mL of BHI broth (Difco) and incubated at 37°C for 24 h in a circulating water bath. One loopful of each 24-h culture was transferred to a separate tube of fresh BHI broth and incubated for an additional 24 h at 37°C. Each overnight culture was then transferred to a separate tube of fresh BHI broth and standardized to an optical density at 600 nm reading (Spectronic GENESYS 2 Spectrophotometer, Spectronic Instruments, Rochester, NY) of 0.5. A 0.33-mL volume from each of the 3 serotype culture tubes was transferred to 100 mL of BHI broth. The mixed inoculum was then incubated for an additional 2.5 h at 37°C in a circulating water bath, yielding a population of ~108 cfu/mL. To verify the initial broth concentration, the inoculum was serially diluted in BPW, spread-plated in duplicate on BHI agar plates, and incubated 24 h at 37°C.

Experimental Design
A complete, crossed factorial design was followed to evaluate the effects of 3 pH levels (4, 7, and 9) and 3 Aw levels (0.84, 0.91, and 0.96) at a constant temperature of 86°F (30°C) on Salmonella growth and survival in poultry litter. To validate the model, an independent experiment was also conducted in litter adjusted to pH 7 and a Aw of 0.91.

Aw Control
Plastic sealable containers (33 x 22.9 x 7.6 cm; Rubbermaid Take Alongs, Rubbermaid Home Products Division, Fairlawn, OH) were used to control Aw. Selected saturated salt solutions corresponding to the target Aw values (Winston and Bates, 1960; Himathongkham et al., 1999) at 30°C (Table 1Go) were prepared, and 150 mL was transferred into separate containers to achieve a target RH in the sealed container and target Aw in the litter. Hardware cloth measuring 27.9 x 21.6 x 1.3 cm was suspended above the saturated salt solution and supported the litter. Containers were then sealed and allowed to reach a vapor phase equilibrium with the salt solution.


View this table:
[in this window]
[in a new window]

  		Table 1. Equilibrium RH [water activity (Aw)] of saturated salt solutions at 30°C
 
Inoculation and Sampling
On the day of each trial, 40-g aliquots of sterile, dried litter per container were aseptically transferred into individual autoclave bags (20.3 x 30.5 cm; Fisherbrand Autoclave Bag, no. 01-830, Fisher Scientific Int.), misted with a plastic sprayer containing 4.5 mL of inoculum to obtain an initial starting population of ~107 cfu/g, and then thoroughly mixed for 1 min. The Aw of the litter was then adjusted to the target level by misting 2 to 5 mL of sterile water onto the litter, mixed, and then aseptically spread across the suspended hardware cloth. Containers were then sealed and placed into a 30°C incubator. Three containers per treatment combination were randomly sampled at scheduled sample times and then 25 g of litter was aseptically transferred to separate sterile filtered stomacher bags (17.8 x 30.5 cm; Fisher Scientific Int.) containing 100 mL of BPW and stomached for 1 min. An additional 125 mL of BPW was added to the homogenized samples, yielding a 1:10 dilution (Beli et al., 2001; Andrews and Hammack, 2003). The bags were agitated for an additional 60 s and then 1 mL of filtered solution was serially diluted in BPW. To enumerate Salmonella populations, diluted samples were spiral-plated (Autoplate 4000, automated spiral plater, Spiral Biotech Inc., Norwood, MA) onto BHI agar plates, incubated at 37°C for 24 h, and the number of colonies was enumerated using an automated colony counting system (ProtoCOL automated colony counter, Synoptics Ltd., Frederick, MD). Suspect colonies were picked, streaked and stabbed onto triple sugar iron agar slants (Difco; Cox et al., 2000), and then incubated at 37°C for 24 h. Positive Salmonella colonies were confirmed by agglutination using poly-O antiserum (Difco; Cox et al., 2000).

The residual litter from each container was used for Aw and pH analysis. Litter Aw was measured according to the manufacturer’s instructions using a Aw meter (Decagon Model CX-3 Water Activity System, Decagon Devices Inc., Pullman, WA). Litter pH was measured as previously described.

Statistical Analysis and Data Fitting
All Salmonella populations were transformed to a base-10 logarithm before analysis. Nonlinear regression was performed on the survival data for each individual Aw and pH treatment combination, using the NLIN procedure of SAS with the Gauss-Newton algorithm (SAS Institute, 1996). Data exhibiting bacterial growth followed by death were fitted using the Churchill model, as described by Membre et al. (1997). A general statistical model for population change over time is log N = µ(t) + {varepsilon}, where µ(t) = the population mean at time t; and {varepsilon} = random error about that mean. The Churchill model assumes that mean population at time t depends on 2 functional components, f1(t) and f2(t), through the expression µ(t) = {f1(t)–1 + f2(t)–1}–1.

These following 2 components are taken as exponential microbial growth and death: f1(t) = K1 exp{{lambda}1t}, f2(t) = K2 exp{–{lambda}2t}, respectively, with the constraint that {lambda}1, {lambda}2 >0. The overall equation for the mean, parameterized by K1, K2, {lambda}1, and {lambda}2 is then µ(t; K1, K2, {lambda}1, {lambda}2) = [1/K1 exp{– {lambda}1t} + 1/K2 exp{{lambda}2t}]–1.

Data exhibiting only bacterial decline were fitted using the general exponential inactivation model for the mean population at time t: µ(t) = K exp{–{lambda}1t}.

General models were then developed from either the Churchill equation (conditions producing both growth and decline) or the general inactivation equation (conditions producing only inactivation) to fit grouped environmental conditions.

Corresponding decimal reduction times (D-values) were calculated as the negative reciprocal of the survivor curve slope obtained by linear regression analysis from the observed and predicted values. Only sampling points along the linear portion of the survivor curve were included in the calculation. For each sample time, the 3 observed replicate sample values were randomly assigned to 1 of 3 data sets, and 3 separate observed D-values along with 1 predicted D-value were calculated for each trial. An ANOVA of the observed D-values across treatments was carried out using the GLM procedure of SAS. This ANOVA involved main and interaction effects for Aw and pH. Subsequent multiple comparisons among means for the 9 treatment combinations were carried out using LSMEANS and t-tests. Agreement between the observed D-values and those predicted by the model was investigated by computing their correlation coefficient (r = 0.998), by checking that the predicted values were within the 95% confidence limits, and by using a paired t-test of the 9 differences. Separate litter containers served as the experimental unit for statistical analysis, and residual effects were used as the error term.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
REFERENCES
 
The behavior of Salmonella in poultry litter as a function of varying pH and Aw treatment combinations is shown in Figures 1Go, 2Go, and 3Go. The Churchill model was successful at fitting data from those treatments demonstrating both bacterial growth and decline; however, a general lack of fit was observed when fitting the model to data showing only bacterial decline. For this reason, the exponential inactivation model was successfully used for fitting those treatment conditions that yielded only bacterial decline.


Figure 1
View larger version (7K):
[in this window]
[in a new window]

  		Figure 1. Observed Salmonella litter population data ({circ}) and predicted values by the Churchill model ({blacktriangleup}) and the exponential inactivation model ({square}) at pH 9 (water activity Aw) 0.96 (panel A), Aw 0.91 (panel B), and Aw 0.84 (panel C).

 

Figure 2
View larger version (7K):
[in this window]
[in a new window]

  		Figure 2. Observed Salmonella litter population data ({circ}) and predicted values by the Churchill model ({blacktriangleup}) and the exponential inactivation model ({square}) at pH 7 (water activity Aw) 0.96 (panel A), Aw 0.91 (panel B), and Aw 0.84 (panel C).

 

Figure 3
View larger version (7K):
[in this window]
[in a new window]

  		Figure 3. Observed Salmonella litter population data ({circ}) and predicted values by the exponential inactivation model ({square}) at pH 4 (water activity Aw) 0.96 (panel A), Aw 0.91 (panel B), and Aw 0.84 (panel C).

 
Target litter conditions of pH 4, 7, and 9 were chosen to represent treated, fresh, and used litter, whereas Aw values of 0.84, 0.91, and 0.96 were chosen to simulate dry, damp, and wet litter and represent a common range of litter Aw values from which positive Salmonella samples have been detected (J. Payne, unpublished data; Opara et al., 1992; Carr et al., 1995; Mallinson et al., 2000; Hayes et al., 2000; de Rezende et al., 2001). Average pH and Aw values for each trial are reported in Table 2Go. A constant incubation temperature of 86°F (30°C) was chosen based on average poultry litter temperatures collected from commercial field data taken over a 2-yr period (J. Payne, unpublished data).


View this table:
[in this window]
[in a new window]

  		Table 2. Average litter pH and water activity (Aw) values1 based on target conditions for each trial
 
Salmonella populations exhibited an initial growth phase (~2 log) with little decline up to 42 d in litter environments of pH 7 and 9 and a Aw value of 0.96 (Figures 1Go and 2Go). This agrees with reports that Salmonella grows at Aw levels ≥0.93 (D’Aoust, 2001). However, at litter conditions of pH 9 and Aw 0.91, slight growth was observed, followed by a steady decline (~4 log at 42 d). As litter Aw levels were reduced to 0.84, populations further declined, with the most drastic reductions (~5 log in 9 h) occurring in low-pH (4) and low-Aw (0.84) environments (Figures 1Go, 2Go, and 3Go). These findings agree with the work of Carr et al. (1995), who found that high Aw values (0.90 to 0.95) in poultry litter were associated with Salmonella-positive broiler flocks, low Aw values (0.79 to 0.84) were associated with Salmonella-negative flocks, and medium Aw values (0.85 to 0.89) indicated an increased risk of contamination. De Rezende et al. (2001) reported similar results in which broiler litter samples having Aw values from 0.90 to 0.95 exhibited the highest Salmonella populations (average of 44.7 cfu/10 g), and litter Aw values ranging from 0.85 to 0.90 produced the second best growth condition (average of 20.7 cfu/10 g). Interestingly, litter samples with Aw values from 0.95 to 1.00 were not associated with high Salmonella populations. These investigators speculated that such conditions are not ideal for Salmonella growth due to a hypotonic environment or the dilution of essential nutrients. In other field studies, Salmonella spp. survived in poultry litter at Aw levels of <0.84; however, the majority of positive samples (74.4%) had Aw values ≥0.90, with most positive samples falling into the 0.90 to 0.95 range (Hayes et al., 2000). These researchers concluded that maintaining poultry litter below 0.84 can aid in controlling Salmonella populations in commercial poultry houses.

Regardless of litter Aw, pH 4 adjusted litter resulted in a rapid decline in Salmonella populations to undetectable levels. At pH 4 and Aw values of 0.96, 0.91, and 0.84, Salmonella populations were below the detection limit (1 log) at 20, 13, and 9 h, respectively (Figure 3Go). At pH 7 conditions, no Salmonella growth was observed in 0.91 and 0.84 Aw environments, with ~5 log reductions in populations at 42 and 12 d, respectively (Figure 2Go). At pH 9 and Aw of 0.84, no Salmonella growth was observed, with ~5 log reduction detected at 10 d (Figure 1Go). Studies have shown that S. Typhimurium and E. coli grow optimally in pH environments from 5 to 9 (Foster, 1993; Small et al., 1994), although Salmonella growth rates generally thrive from pH 6.5 to 7.5 (Chung and Goepfert, 1970; D’Aoust, 1989). Others have reported that the pH growth range for Salmonella falls between 3.6 and 9.5, with optimal growth at near neutral pH (D’Aoust, 2001). In a study in which Salmonella-positive drag swabs were isolated from broiler house litter, mean poultry litter pH values were 8.1 ± 0.1 (Opara et al., 1992). However, another study has shown that the reduction of litter pH to a more acidic level (pH 4) resulted in a decline in microbial populations, including E. coli, Salmonella, and Clostridium, to below detectable limits (Hardin and Roney, 1989). Our findings agree with these reports and suggest that if poultry litter pH can be lowered to pH 4 or below, then Salmonella populations may be drastically reduced or possibly eliminated in commercial poultry litter.

A 2-way factorial analysis of pH and Aw effects on observed litter Salmonella population D-values was then conducted and yielded a significant pH and Aw interaction (Table 3Go). The D-values were significantly longer under litter conditions of pH 9 and 7 and a Aw of 0.96 than all other experimental conditions (P < 0.05). The shortest D-values occurred in pH 4 environments. Under these acidic litter conditions, a significant increase in the Salmonella death rate (~200-fold) occurred compared with litter environments of pH 9 and 7 at a Aw of 0.96 (P < 0.05).


View this table:
[in this window]
[in a new window]

  		Table 3. Observed and predicted D-values (d) of Salmonella litter populations as affected by litter pH and water activity (Aw) effects
 
Predicted D-values by the Churchill and inactivation models were compared with the average observed D-values for each trial and are shown in Table 3Go. For all comparisons, no significant differences were observed, indicating the plausibility of the models in describing the response of Salmonella to pH and Aw (P < 0.05).

Due to extreme differences in response detected across the 9 treatments and the significant interaction between pH and Aw, a general model was not successfully fitted for the 9 combined conditions. However, generalized models for bacterial growth and death under specified pH environments were successfully developed to predict Salmonella behavior in litter over time. The Churchill equation was effectively used to develop a general Salmonella growth and inactivation model, fitting data from combined pH 7 and 9 experimental conditions. The parameter {lambda}2 was set to be dependent on pH and Aw levels, with d0 representing the intercept and d1 and d2 representing pH and Aw parameter estimates, respectively. At time 0, the relationship between K1 and K2, as described by Membre et al. (1997), was as follows: 1/K1 = (log N0)–1 –1/K2.

Log N0 was treated as a known parameter and assigned a Salmonella log population count of 7 at time 0. The expressions used for the model are listed below:


Formula


Formula

where K1 = the growth phase and {lambda}2 = the death phase. These parameters were then substituted into the Churchill equation and fitted to the experimental data.

Although the generalized Churchill model successfully fit data from experimental conditions favoring growth followed by death, a lack of fit was observed for data from environments resulting in bacterial decline only. The drastic reductions in Salmonella populations (~5 log in <21h to below detection limits) observed at pH 4 required the development of a general inactivation model specific to that environment. Similarly, a general bacterial inactivation model was required to fit pH 7 and 9 litter conditions yielding only bacterial decline (i.e., pH 7 at Aw 0.84 and 0.91 and pH 9 at Aw 0.84). Because both pH 7 and 9 environments resulted in similar rates of Salmonella inactivation, a general model was successfully developed to describe both of these pH environments. Both exponential inactivation models follow the form previously described.

The equation used for the pH 4 inactivation model was as follows: {lambda} = b0 + b1 x Aw. The equation used for the pH 7 and 9 inactivation model was as follows: {lambda} = b0 + b1 x pH + b2 x Aw. The {lambda} parameter ({lambda}) in the pH 4 inactivation model was set to be dependent on Aw levels, with b0 and b1 representing the intercept and slope for Aw, respectively. The pH 7 and 9 inactivation model {lambda} was dependent on pH and Aw, with b0, b1, and b2 representing the intercept and slopes for pH and Aw, respectively. Figures 4Go, 5Go, and 6Go represent observed and predicted log10 Salmonella populations for the pH 7 and 9 growth and inactivation (Churchill), pH 7 and 9 inactivation, and pH 4 inactivation general models, respectively. Each general model accounts for the 3 Aw levels tested in the experimental design in 1 single step.


Figure 4
View larger version (7K):
[in this window]
[in a new window]

  		Figure 4. Generalized pH 7 and 9 growth and inactivation model with water activity (Aw) and pH effects. Observed Salmonella litter population data (open symbols) and corresponding predicted values (closed symbols) by the Churchill model at pH 7, Aw 0.96 ({diamond}); pH 9, Aw 0.96 ({triangleup}); and pH 9, Aw 0.91 ({circ}).

 

Figure 5
View larger version (8K):
[in this window]
[in a new window]

  		Figure 5. Generalized pH 7 and 9 inactivation model with water activity (Aw) and pH effects. Observed Salmonella litter population data (open symbols) and corresponding predicted values (closed symbols) by the exponential inactivation model at pH 7, Aw 0.91 ({diamond}); pH 7, Aw 0.84 ({triangleup}); and pH 9, Aw 0.84 ({circ}).

 

Figure 6
View larger version (6K):
[in this window]
[in a new window]

  		Figure 6. Generalized pH 4 inactivation model with water activity (Aw) effects. Observed Salmonella litter population data (open symbols) and corresponding predicted values (closed symbols) by the exponential inactivation model at Aw 0.96 ({diamond}); 0.91 ({triangleup}); and 0.84 ({circ}).

 
The Churchill general model showed significant Aw effects on the Salmonella decline rate ({lambda}2), with a d2 parameter estimate value of –0.50 and a 95% confidence interval of –0.64 to –0.37 (P < 0.05); however, pH did not affect Salmonella death. These Aw effects are evident in Figure 4Go; Salmonella populations declined as litter Aw was lowered from 0.96 to 0.91. Significant differences in D-values (Table 3Go) between pH 9 and 7 and a Aw of 0.96 and pH 7 and Aw 0.91 litter environments further confirm this observation (P < 0.05). The pH 7 and 9 general inactivation model (Figure 5Go) showed significant pH and Aw effects on Salmonella decline or {lambda} with b1 and b2 parameter estimates of 0.02 and –1.45 and 95% confidence limits of 0.001 to 0.04 and –1.85 to –1.06, respectively. Significant Aw effects on {lambda} were found using the pH 4 general inactivation model (Figure 6Go), with a b1 parameter estimate value of –0.36 and 95% confidence interval of –0.50 to –0.21.

Parameter estimates from each of the 3 general models were used to calculate {lambda} values for the 9 litter pH and Aw treatment combinations and are listed in Table 4Go. These values can be used to describe the log Salmonella decline rate per unit time. Litter pH 4 {lambda} values were converted from unit time hours to days for ease of comparison. For pH 4, 7, and 9 environments, {lambda} values were lower at Aw 0.96 (2.134, 0.005, and 0.036) and 0.91 (2.563, 0.039, and 0.029) and higher at Aw 0.84 (3.165, 0.141, 0.190), respectively, showing that Salmonella decline rates increased as Aw decreased to 0.84. Litter pH 4 at Aw 0.96, 0.91, and 0.84 conditions yielded the highest {lambda} values (2.134, 2.563, and 3.165), respectively.


View this table:
[in this window]
[in a new window]

  		Table 4. Lambda values describing Salmonella decline rates1 from generalized models under specified pH and water activity (Aw) litter conditions
 
The adequacy of each model to predict Salmonella populations can be assessed graphically by comparing observed values against corresponding predicted values, thus testing the overall reliability of the model (Giffel and Zwietering, 1999). Statistical comparison by calculating the MS error (MSE) and regression coefficient are other tools for assessing the fit of the model. Mean square error is a measure of the variability remaining and not accounted for by Aw and pH effects, whereas regression coefficient measures the fraction of the variation about the mean that is explained by the model. The lower the MSE and the higher the regression coefficient values, the more adequate the model describes the data (Giffel and Zwietering, 1999). Mean square error and regression coefficient values for both individual and generalized models are listed in Table 5Go. Individual Churchill models fitting data for all 3 growth and death conditions (pH 7 and Aw 0.96; pH 9 and Aw 0.91 and 0.96) resulted in the lowest MSE values. For these 3 trials, MSE values were lower than those reported by Tamagnini et al. (2004), who found good fits using the Churchill model to describe S. Typhimurium behavior in goat cheese stored at 5, 15, and 25°C. This comparison suggests that the Churchill model accurately describes the growth and death data. Although MSE values were higher for individual inactivation models than individual Churchill models, all models resulted in a high regression coefficient value (≥0.96), which indicates a good fit. When comparing the generalized models, the Churchill model produced the lowest MSE value; however, all generalized models resulted in high regression coefficient values (≥0.97).


View this table:
[in this window]
[in a new window]

  		Table 5. Mean square error (MSE) and regression coefficient values for individual and general models describing pH and water activity (Aw) effects on Salmonella litter populations
 
To further validate the generalized model, another set of data from a separate independent study conducted at litter pH 7 and Aw 0.91 conditions was compared with the predicted values of the pH 7 and 9 general inactivation model and are shown in Figure 7Go. Visual inspection of the graph indicates that the general model provided a good fit to the observed data.


Figure 7
View larger version (6K):
[in this window]
[in a new window]

  		Figure 7. External validation data set (open symbols) and corresponding predicted values (closed symbols) by the generalized pH 7 and 9 inactivation model of Salmonella litter populations at pH 7, water activity 0.9 ({diamond}).

 
Predictive modeling has been used in the past to describe the effect of environmental factors and their interactive effects on the microbial behavior of foodborne pathogens (Gibson et al., 1988; Buchanan and Phillips, 1990; Wijtzes et al., 1993; Linton et al., 1995; Koutsoumanis et al., 1999). The USDA-based Microbial Pathogen Computer Modeling Program can be used as a valuable tool for food companies when supporting HACCP requirements and is now available in a series of programs that can be run on personal computers (Whiting and Buchanan, 2001). A food manufacturing company can use these programs as a tool for estimating the influence of a limiting agent or the combination of agents on pathogens and evaluating potential problems that might occur during processing. Pathogen modeling programs are useful in supporting hazard analysis, developing critical limits, and predicting the effectiveness of corrective actions taken.

Some members of the poultry industry are currently evaluating the feasibility of developing pathogen-reduction programs at the growout phase in an effort to proactively address the potential for future implementation of preharvest HACCP-like programs. Should pathogen control be mandated on the farm, both integrators and growers will be challenged to identify effective pathogen control measures for use during growout. Studies supporting model development for pathogen control during live production have application toward a "farm to fork" approach for reducing the spread of foodborne pathogens to humans through contaminated food products and the environment.

Pathogen models can serve as a useful tool for assessing the risk of contamination or contamination levels on Salmonella-positive farms. Prior knowledge of potentially high or low Salmonella contamination populations in the growout environment can be useful when making farm management and processing decisions affecting food safety. For example, when selecting the order of flock arrival to processing plants, Salmonella-positive flocks with the greatest potential for high Salmonella populations could be the last in order for processing on a given day. Our results indicate that pH conditions of 7 and 9 at a Aw of 0.96 are favorable for Salmonella growth and survival up to market age (42 d). Decisions could be made based on model predictions while avoiding laborious and time-constraining sampling and analysis for Salmonella populations that can often take days to yield results. This practice could reduce cross-contamination risks at the processing plant and assure compliance with USDA Salmonella performance standards.

Models have application not only for processing aspects but also for live production. Newly hatched chicks are more susceptible to Salmonella infection than more mature birds (Gast, 1997). As few as 5 cells of Salmonella have been shown to infect chicks (Milner and Shaffer, 1952), and this number may even be lower if the birds are stressed (Arakawa et al., 1992). Once infected, these birds may excrete fecal concentrations of up to 109 Salmonella/g of feces for up to a 2-wk duration (Bailey, 1987). Chick mortality has been observed to reach its peak at 3 to 7 d (Gast, 1997). In some instances, new litter has been shown to be contaminated with Salmonella before bird placement (Kumar et al., 1971, Simmons and Byrnes, 1972; Bhatia et al., 1979). Under these circumstances, a statistical model can be useful for determining target environmental conditions that are unfavorable for Salmonella proliferation and for estimating the amount of time needed to achieve significant population reductions to a safe level under given conditions. Our results showed that achieving a low pH and Aw litter environment led to the fastest reduction in Salmonella populations when compared with higher pH and Aw litter conditions. Best management practices can then be linked to attaining these targeted environmental conditions.

Controlling RH inside the house and in the litter is an important control strategy for reducing pathogens, ammonia fumes, and parasites such as coccidia in the bird’s environment (Zander et al., 1997). Equilibrium RH, which is the RH of the surrounding air, is equivalent to Aw (ERH% = Aw x 100). The measurement of Aw can thus be achieved by measuring the ERH of the atmosphere surrounding the sample (Adams and Moss, 1995). Water activity measures available or free water, whereas moisture content measurements include both bound (unavailable) and unbound (free) water (Opara et al., 1992). Mallinson et al. (1998) reported that when litter RH values are <92%, Salmonella populations decline. According to Wabeck (1998), Salmonella can be controlled with a reduced litter Aw <0.80 or when litter pH is maintained <4.0.

Proper ventilation practices are not only critical to cooling birds but are also a key management tool used to remove excess moisture from the broiler house and to maintain a certain degree of dryness in the litter. Valentine (1964) found that both ammonia and RH levels were reduced as the rates of air exchange inside well-insulated test pens (8 x 14 ft. or 2.44 x 4.27 m floor area) increased. Mallinson et al. (2000) reported that low broiler litter surface airflow rates (<15.6 m/min or 51 ft/min) were related to increased litter Salmonella populations (1.63 cfu/10 g) compared with higher airflow rates (>15.6 m/ min or 51 ft/min) and decreased Salmonella populations (<1.33 cfu/10 g). These low airflow rates were associated with higher litter moisture content (41.5 vs. 30.9%) and Aw levels (0.91 vs. 0.89) compared with high airflow rates, respectively. In another study, litter areas exposed to higher ventilation rates (18.3m/min or >60 ft/min) in commercial broiler houses were found to be significantly related to lower moisture content (28 vs. 41.2%), Aw (0.84 vs. 0.91), and E. coli populations (8.2 x 105 vs. 5.4 x 106 cfu/10 g) when compared with lower ventilation rates (<18.3m/min or 60 ft/min ), respectively (de Rezende et al., 2001). In the same study, reduced Aw (<0.89) and moisture content (<35%) levels corresponded to reduced Salmonella populations (20.7 and 5.3 cfu/10 g, respectively). A low incidence of Salmonella from litter samples was observed from houses that were tunnel-ventilated (39.9 to 60.3 m/min or 131 to 198 ft/min). This observation suggests that tunnel ventilation could be effective at reducing moisture from the litter, which would in turn reduce Salmonella populations. These investigators recommended maintaining a modest and uniform ventilation rate of 100 to 150 ft/min (30.5 to 45.7 m/m) over the litter to maintain a drier litter environment.

Litter treatments are commonly used in poultry houses to reduce harmful ammonia emissions, but they may also be used to reduce litter pathogens by lowering litter pH. Pope and Cherry (2000) reported significant declines in litter pH and ammonia levels along with reduced total aerobic bacteria and E. coli populations in litter treated with a NaHSO4 product as compared with nontreated houses. Payne et al. (2002) have also shown that lowering litter pH to 2.68 and 3.48 using a H2SO4 and NaHSO4 litter treatment product significantly reduced Salmonella populations by 1.04 and 1.30 log cfu/mL, respectively. Litter treatments are commonly applied before bird placement, thereby reducing litter pH during the first week of rearing when the birds are the most susceptible to pathogen invasion. Our data suggests that by reducing litter pH to 4, Salmonella populations can be reduced below detectable limits within 20 h or less when litter is previously contaminated with high populations of Salmonella (~107). Because pH levels have been shown to rise to near neutral levels 1 wk postapplication of a litter treatment (Pope and Cherry, 2000), reapplication may be required to maintain acidic litter conditions.

It should be noted that predictive models have limitations and should not be relied upon by themselves for determining the safety of foods and other production or processing systems. Independent microbial laboratory studies are still required for validation purposes when determining the growth and survival of pathogens. Moreover, most models do not include all of the growth factors influencing the growth kinetics of the organism under investigation. Other factors, such as the presence of bacteriocins or competitive exclusion effects from other organisms, are often not factored into the model. In many cases, the behavior of a specific pathogen in a specific growth media or food product is investigated and thus the results should only be interpreted for those specified conditions. For this study, results should only be interpreted for Salmonella behavior in presterilized and inoculated used poultry litter and should not be mistaken for mimicking the live production setting. However, our model offers an understanding of how Salmonella behaves under various environmental growth conditions in poultry litter and can be used as a tool for developing control strategies for use during preharvest production.

The data suggest that by lowering litter pH or Aw levels to 4 or 0.84, respectively, and through proper farm management practices, Salmonella populations can be reduced to below detectable limits. The multiple hurdle concept can be successfully implemented by combining the interactive effects of low pH and Aw for the control of Salmonella growth and survival. By reducing Salmonella in the bird’s environment through a multistep intervention program, reductions in prevalence and populations on birds entering processing plants from contaminated flocks should be observed, thus reducing the cross-contamination potential and true risk to humans. Pathogen models, such as those described in this study, have potential application to aid in establishing on-farm food safety programs by predicting pathogen populations under varying environmental conditions.

Received for publication July 13, 2006. Accepted for publication August 22, 2006.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 REFERENCES
 
Adams, M. R., and M. O. Moss. 1995. Pages 18–203 in Food Microbiology. RSC, Cambridge, UK.

Andrews, W. H., and T. S. Hammack. 2003. Salmonella. Bacteriological Analytical Manual Online. http://vm.cfsan.fda.gov/~ebam/bam-5.html Accessed Aug. 2003.

Arakawa, A., T. Fukaton, E. Baba, L. R. McDougald, J. S. Bailey, and L. C. Blankenship. 1992. Influence of coccidiosis on colonization in broiler chickens under floor pen conditions. Poult. Sci. 71:59–63.[ISI][Medline]

Bailey, J. S. 1987. Factors affecting microbial competitive exclusion in poultry. Food Technol. 41:88–92.

Bains, B. S., and M. A. MacKenzie. 1974. Transmission of Salmonella through an integrated poultry operation. Poult. Sci. 53:1114–1118.[ISI][Medline]

Beli, E., A. Telo, and E. Duraku. 2001. Salmonella serotypes isolated from turkey meat in Albania. Int. J. Food Microbiol. 63:165–167.[ISI][Medline]

Bender, F. E., and E. T. Mallinson. 1991. Healthy birds are lower cost birds. Broiler Ind. 54:62–64.

Bhatia, T. R. S., G. D. McNabb, H. Wyman, and G. P. S. Nayar. 1979. Salmonella isolation from litter as an indicator of flock infection and carcass contamination. Avian Dis. 23:838–847.[ISI][Medline]

Bryan, F. L., and M. P. Doyle. 1995. Health risks and consequences of Salmonella and Campylobacter jejuni in raw poultry. J. Food Prot. 58:326–344.

Bryan, F. L., M. J. Fanelli, and H. Riemann. 1979. Salmonella infections. Pages 74–130 in Foodborne Infections and Intoxications. H. Riemann and F. L. Bryan, ed. Acad. Press Inc., London, UK.

Buchanan, R. L., and J. G. Phillips. 1990. Response surface model for predicting the effects of temperature, pH, sodium chloride content, sodium nitrite concentration and atmosphere on the growth of Listeria monocytogenes. J. Food Prot. 53:370–376.

Byrd, J. A., D. E. Corrier, J. R. DeLoach, and D. J. Nisbet. 1997. Comparison of drag-swab environmental protocols for the isolation of Salmonella in poultry houses. Avian Dis. 41:709–713.[ISI][Medline]

Carr, L. E., E. T. Mallinson, C. R. Tate, R. G. Miller, E. Russek-Cohen, L. E. Stewart, O. O. Opara, and S. W. Joseph. 1995. Prevalence of Salmonella in broiler flocks: Effect of litter and water activity, house construction, and watering devices. Avian Dis. 39:39–44.[ISI][Medline]

Chung, K. C., and J. M. Goepfert. 1970. Growth of Salmonella at low pH. J. Food Sci. 35:326–328.

Churchill, S. W., and R. Usagi. 1972. A general expression for the correlation of rates of transfer and other phenomena. AIChE J. 7:38–41.

Corrier, D. E., J. A. Byrd, B. M. Hargis, M. E. Hume, R. H. Bailey, and L. H. Stanker. 1999. Presence of Salmonella in the crop and ceca of broiler chickens before and after preslaughter feed withdrawal. Poult. Sci. 78:45–49.[Abstract/Free Full Text]

Cox, N. A., N. J. Stern, S. E. Craven, M. E. Berrang, and M. T. Musgrove. 2000. Prevalence of Campylobacter and Salmonella in the cecal droppings of turkeys during production. J. Appl. Poult. Res. 9:542–545.[Abstract/Free Full Text]

D’Aoust, J.-Y. 1989. Salmonella. Pages 327–445 in Foodborne Bacterial Pathogens. M. P. Doyle, ed. Marcel Dekker Inc., New York, NY.

D’Aoust, J.-Y. 2001. Salmonella. Pages 163–191 in Guide to Food-borne Pathogens. R. G. Labbe and S. Garcia, ed. John Wiley and Sons Inc., New York, NY.

de Rezende, C. L., E. T. Mallinson, N. L. Tablante, R. Morales, and A. Park. 2001. Effect of dry litter and airflow in reducing Salmonella and Escherichia coli populations in the broiler production environment. J. Appl. Poult. Res. 10:245–251.[Abstract/Free Full Text]

Farkas, J. 2001. Physical methods of food preservation. Pages 497–519 in Food Microbiology: Fundamentals and Frontiers. M. P. Doyle, L. R. Beuchat, and T. J. Montville, ed. ASM Press. Washington, DC.

Foster, J. W. 1993. The acid tolerance response of Salmonella Typhimurium involves transient synthesis of key acid shock proteins. J. Bacteriol. 175:1981–1987.[Abstract/Free Full Text]

Gast, R. K. 1997. Salmonella infections. Paratyphoid infections. Pages 97–121 in Diseases of Poultry. W. B. Calnek, ed. Iowa State Univ. Press, Ames.

Gibson, A. M., N. Bratchell, and T. A. Roberts. 1988. Predicting microbial growth: Growth responses of salmonellae in a laboratory medium as affected by pH, sodium chloride, and storage temperature. Int. J. Food Microbiol. 6:155–178.[ISI][Medline]

Giffel, M. C., and M. H. Zwietering. 1999. Validation of predictive models describing the growth of Listeria monocytogenes. Int. J. Food Microbiol. 46:135–149.[ISI][Medline]

Hardin, B. E., and C. S. Roney. 1989. Effects of pH on selected bacteria. Alabama Department of Agriculture and Industry Report.

Hayes, J. R., L. E. Carr, E. T. Mallinson, L. W. Douglass, and S. W. Joseph. 2000. Characterization of the contribution of water activity and moisture content to the population distribution of Salmonella spp. in commercial poultry houses. Poult. Sci. 79:1557–1561.[Abstract/Free Full Text]

Himathongkham, S., S. Nuanualsuwan, and H. Riemann. 1999. Survival of Salmonella Enteritidis and Salmonella Typhimurium in chicken manure at different levels of water activity. FEMS Microbiol. Lett. 172:159–163.[ISI][Medline]

Hoszowski, A., A. D. E. Fraser, B. W. Brooks, and E. M. Riche. 1996. Rapid detection and enumeration of Salmonella in chicken carcass rinses using filtration, enrichment and colony blot immunoassay. Int. J. Food Microbiol. 28:341–350.[ISI][Medline]

Jones, F. T., R. C. Axtell, D. V. Rives, S. E. Scheideler, F. R. Tarver Jr., R. L. Walker, and M. J. Wineland. 1991. A survey of Salmonella contamination in modern broiler production. J. Food Prot. 54:502–507.

Kiessling, C. R., J. H. Cutting, M. Loftis, W. M. Kiessling, A. R. Datta, and J. N. Sofos. 2002. Antimicrobial resistance of food-related Salmonella isolates, 1999–2000. J. Food Prot. 65:603–608.[ISI][Medline]

Koutsoumanis, K., K. Lambropoulou, and G.-J. E. Nychas. 1999. A predictive model for the non-thermal inactivation of Salmonella Enteritidis in a food model system supplemented with a natural antimicrobial. Int. J. Food Microbiol. 49:63–74.[ISI][Medline]

Kumar, J., M. York, J. McDowell, and B. Pomeroy. 1971. Dynamics of Salmonella infection in fryer roaster turkeys. Avian Dis. 15:221–232.[ISI][Medline]

Lahellec, C., and P. Colin. 1985. Relationship between serotypes of Salmonella from hatcheries and rearing farms and those from processed poultry carcasses. Br. Poult. Sci. 26:179–186.[ISI][Medline]

Leistner, L. 2000. Basic aspects of food preservation by hurdle technology. Int. J. Food Microbiol. 55:181–186.[ISI][Medline]

Linton, R. H., W. H. Carter, M. D. Pierson, and C. R. Hackney. 1995. Use of modified Gompertz equation to model nonlinear survival curves for Listeria monocytogenes. J. Food Prot. 60:1497–1501.

Mallinson, E. T., C. E. de Rezende, N. L. Tablante, L. E. Carr, and S. W. Joseph. 2000. A management technique to identify prime locations of Salmonella contamination on broiler and layer farms. J. Appl. Poult. Res. 9:364–370.[Abstract/Free Full Text]

Mallinson, E. T., S. W. Joseph, and L. E. Carr. 1998. Salmonella’s Achilles’ heel. Broiler Ind. 61:22–32.

McMeekin, T. A., J. N. Olley, T. Ross, and D. A. Ratkowsky. 1993. Pages 69–70 in Predictive Microbiology: Theory and Application. Research Studies Press Ltd. Somerset, UK.

Membre, J. M., J. Thurette, and M. Catteau. 1997. Modelling the growth, survival and death of Listeria monocytogenes. J. Appl. Microbiol. 82:345–350.[Medline]

Milner, K. C., and M. F. Shaffer. 1952. Bacteriological studies of experimental Salmonella infections in chicks. J. Infect. Dis. 90:81–96.[ISI][Medline]

Opara, O. O., L. E. Carr, E. Russek-Cohen, C. R. Tate, E. T. Mallinson, R. G. Miller, L. E. Stewart, R. W. Johnston, and S. W. Joseph. 1992. Correlation of water activity and other environmental conditions with repeated detection of Salmonella contamination on poultry farms. Avian Dis. 36:664–671.[ISI][Medline]

Payne, J. B., E. C. Kroger, and S. E. Watkins. 2002. Evaluation of litter treatments on Salmonella recovery from poultry litter. J. Appl. Poult. Res. 11:239–243.[Abstract/Free Full Text]

Pope, M. J., and T. E. Cherry. 2000. An evaluation of the presence of pathogens on broilers raised on poultry litter treatment-treated litter. Poult. Sci. 79:1351–1355.[Abstract/Free Full Text]

SAS Institute. 1996. SAS/STAT User’s guide. Version 6, 4th ed, vol. 2. SAS Inst. Inc., Cary, NC.

Silliker, J. H. 1980. Status of Salmonella—Ten years later. J. Food Prot. 43:307–313.

Simmons, G., and R. Byrnes. 1972. The origin of salmonellas in chickens and chicken carcasses. Aust. Vet. J. 48:186–189.[ISI][Medline]

Small, P., D. Blankenship, D. Welty, E. Zinser, and J. L. Slonczewski. 1994. Acid and base resistance in Escherichia coli and Shigella flexneri: Role of rpoS and growth pH. J. Bacteriol. 176:1729–1737.[Abstract/Free Full Text]

Tamagnini, L. M., G. B. de Sousa, R. D. Gonzalez, J. Revelli, and C. E. Budde. 2004. Behavior of Yersinia enterocolitica and Salmonella Typhimurium in Crottin goat’s cheese. Int. J. Food Microbiol. 99:129–134.

Tauxe, R. V. 1991. Salmonella: A postmodern pathogen. J. Food Prot. 54:563–568.

Trampel, D. W., R. J. Hasiak, L. J. Hoffman, and M. C. Debey. 2000. Recovery of Salmonella from water, equipment, and carcasses in turkey processing plants. J. Appl. Poult. Res. 9:29–34.[Abstract/Free Full Text]

Turnbull, P. C. B., and G. H. Snoeyenbos. 1973. The roles of ammonia, water activity, and pH in the salmonellacidal effect of long-used poultry litter. Avian Dis. 17:72–86.[ISI][Medline]

US Department of Agriculture Food Safety and Inspection Service. 1995. Nationwide broiler chicken microbiological baseline data collection program, July 1994–June 1995 http://www.fsis.usda.gov/Science/Baseline_Data/index.asp Accessed April 2005.

US Department of Agriculture Food Safety and Inspection Service. 1998. Farm-to-table safety system; Salmonella enteritidis contamination, control and reduction. Fed. Regist. 63:27502–27511.

U.S. Department of Agriculture Food Safety and Inspection Service. 2003. Progress report on Salmonella testing of raw meat and poultry products, 1998–2002. http://www.fsis.usda.gov/OPHS/haccp/salm5year.htm Accessed April, 2005.

Valentine, H. 1964. A study of the effect of different ventilation rates on the ammonia concentrations in the atmosphere of broiler houses. Br. Poult. Sci. 5:149–159.

Wabeck, C. J. 1998. HACCP down on the farm. Broiler Ind. 61:36–39.

Whiting, R. C., and R. L. Buchanan. 2001. Food microbiology fundamentals and frontiers. Pages 728–739 in Predictive Modeling. M. P. Doyle, L. R. Beuchat, and T. J. Montville, ed. ASM Press, Washington, DC.

Wijtzes, T., P. J. McClure, M. H. Zwietering, and T. A. Roberts. 1993. Modelling bacterial growth of Listeria monocytogenes as a function of water activity, pH and temperature. Int. J. Food Microbiol. 18:139–149.[ISI][Medline]

Winston, P. W., and D. H. Bates. 1960. Saturated solutions for the control of humidity in biological research. Ecology 41:232–237.[ISI]

Zander, D. V., A. J. Bermudez, and E. T. Mallinson. 1997. Principles of disease prevention: Diagnosis and control. Pages 1–45 in Diseases of Poultry. W. B. Calnek, ed. Iowa State University Press, Ames.





This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Payne, J. B.
Right arrow Articles by Sheldon, B. W.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Payne, J. B.
Right arrow Articles by Sheldon, B. W.

HOME HELP