Growth and inactivation of Salmonella at low refrigerated storage temperatures and thermal inactivation on raw chicken meat and laboratory media: Mixed effect meta-analysis

Hanan Smadi; Jan M. Sargeant; Harry S. Shannon; Parminder Raina

doi:10.1016/j.jegh.2012.12.001

Download article (PDF)

Next Article In Issue>

Volume 2, Issue 4, December 2012, Pages 165 - 179

Growth and inactivation of Salmonella at low refrigerated storage temperatures and thermal inactivation on raw chicken meat and laboratory media: Mixed effect meta-analysis

Authors

Hanan Smadi^a^{, b}^{, *}^,smadica@yahoo.ca, Jan M. Sargeant^b^{, c}, Harry S. Shannon^a, Parminder Raina^a

^aDepartment of Clinical Epidemiology and Biostatistics, McMaster University, Hamilton, Ontario, Canada L8S 4K1

^bCentre for Public Health and Zoonoses, University of Guelph, Guelph, Ontario, Canada N1G 2W1

^cDepartment of Population Medicine, Ontario Veterinary College, Guelph, Ontario, Canada N1G 2W1

^*Corresponding author at: Centre for Public Health and Zoonoses, University of Guelph, Guelph, Ontario, Canada N1G 2W1.

Corresponding Author

Hanan Smadismadica@yahoo.ca

Received 22 October 2012, Revised 1 December 2012, Accepted 3 December 2012, Available Online 27 December 2012.

DOI: 10.1016/j.jegh.2012.12.001 How to use a DOI?
Keywords: Refrigeration; Thermal inactivation; Salmonella; Broiler chicken; Mixed effect; Meta-analysis
Abstract: Growth and inactivation regression equations were developed to describe the effects of temperature on Salmonella concentration on chicken meat for refrigerated temperatures (⩽10 °C) and for thermal treatment temperatures (55–70 °C). The main objectives were: (i) to compare Salmonella growth/inactivation in chicken meat versus laboratory media; (ii) to create regression equations to estimate Salmonella growth in chicken meat that can be used in quantitative risk assessment (QRA) modeling; and (iii) to create regression equations to estimate D-values needed to inactivate Salmonella in chicken meat. A systematic approach was used to identify the articles, critically appraise them, and pool outcomes across studies. Growth represented in density (Log₁₀ CFU/g) and D-values (min) as a function of temperature were modeled using hierarchical mixed effects regression models. The current meta-analysis analysis found a significant difference (P ⩽ 0.05) between the two matrices – chicken meat and laboratory media – for both growth at refrigerated temperatures and inactivation by thermal treatment. Growth and inactivation were significantly influenced by temperature after controlling for other variables; however, no consistent pattern in growth was found. Validation of growth and inactivation equations against data not used in their development is needed.
Copyright: © 2012 Ministry of Health, Saudi Arabia. Published by Elsevier Ltd.
Open Access: This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)

1. Introduction

Salmonellosis is one of the main bacterial food-borne illnesses in Canada and worldwide [1]. In humans, salmonellosis is primarily a disease confined to the gastrointestinal tract, but may cause serious extra-intestinal tract disease, especially in the very young, the aged and those that are immunologically compromised [2]. The symptoms of salmonellosis include nausea, vomiting, abdominal cramps, fever and headaches with a duration that ranges from days to weeks [3]. It is primarily transmitted from infected carrier animals to humans through contaminated food [4]. Meat in general and poultry in particular are the most common sources of food-borne illness by Salmonella [5]. Contamination of poultry can occur at multiple steps along the food chain, including production, processing, distribution, retail marketing, and handling/preparation [6].

Microbial quantitative risk assessment (QRA) has been incorporated in the decision-making process of the Codex Alimentarius Commission (CAC) to manage public health risks associated with microbial hazards [7]. Mathematical models for use in QRA to predict Salmonella growth as a function of temperature, pH and water activity are available in the literature [8–10]. However, most predictive models designed for Salmonella growth have used laboratory media, such as brain/heart infusion broth and not actual chicken to develop prediction models for Salmonella growth [11]. This might overestimate growth owing to the absence of competitive micro flora usually available in raw chicken meat [12].

Effects of environmental conditions such as temperature and its impact on growth kinetics of Salmonella in real food products have been studied less extensively. To address this gap, Oscar studied the impact of temperature on Salmonella growth on cooked chicken breast [13–16], and raw chicken [12,17]. For raw chicken, the studied temperature range was 10–40 °C [12,17]. Considering the broad range of temperature where Salmonella can grow (5–47 °C) [18], growth at temperatures ⩽10 °C in raw chicken still needs to be investigated. As far as this research is concerned, meta-analysis to pool the available data on the effect of temperatures ⩽10 °C has not been conducted using data based on real chicken products or laboratory media.

Furthermore, an important contributing factor that leads to salmonellosis is inadequate temperature/time exposure during the cooking process to kill the pathogenic bacteria [19]. Insufficient cooking has been identified as one of the most important factors contributing to food-borne disease in Canada [20]. As a result, cooking is considered to be a primary means of eliminating pathogens from contaminated meat products and hence serves as a protective method for preventing food-borne illnesses [21]. Previous researchers have conducted thermal inactivation studies of Salmonella spp. in aqueous media and foods [18]. However, few researchers have addressed the question of whether Salmonella inactivation in food products such as beef, pork, turkey, or chicken is the same as inactivation in laboratory media. The common approach is to use the estimated D-values, which is the time required at a given temperature to reduce the number of pathogenic bacteria by 90% [22] during heat treatment, in laboratory media and apply these values to food products. Such estimation might underestimate risk as the bacteria attached to meat tissues may be more heat resistant than bacteria suspended in a liquid medium [23]. Therefore, it is important to evaluate the thermal inactivation of individual pathogens in food products.

The objectives of this paper were: (i) to compare Salmonella density (Log₁₀ colony forming units [CFU]/g) and inactivation (D-values) from primary studies available to investigate Salmonella growth/inactivation in chicken meat with that in laboratory media using a meta-analysis approach, and if there is a significant difference between the two matrices; then (ii) to conduct meta-analysis of chicken data to estimate Salmonella growth at refrigerated storage conditions below 10 °C and to create a mathematical equation that could be used in QRA modeling to estimate Salmonella growth at this temperature range; and (iii) to conduct meta-analysis to create a mathematical model for Salmonella inactivation using D-values in chicken products. The meta-analysis approach combines data from a number of individual studies to produce a more precise estimate of the summary outcome [24].

2. Materials and methods

The review carried out the following steps: a comprehensive literature search to identify all potentially relevant research, relevance screening of abstracts identified by the search, full text screening and quality assessment of relevant abstracts, and data extraction. Mixed-effect regression models using SAS PROC MIXED (version 9.1.3 SAS) were conducted on extracted data to predict growth (Log₁₀ CFU) and Log₁₀ D-values as a function of the characteristics of individual studies.

2.1. Literature search

The identification of potentially relevant research began by compiling a comprehensive list of search terms (Table 1). Experimental designs and observational studies were eligible for inclusion to allow for the investigation of different study designs, i.e., methodological heterogeneity. The search terms were entered into six electronic databases to identify abstracts published between January 1960 and 2008: MEDLINE, PubMed, EMBASE, The AGRICultural OnLine Access (AGRICOLA), INGENTA, and ISI Web of Knowledge. The search was limited to words in the title or abstracts and an English language limit was imposed. In addition, hand searching was completed using the search terms presented in Table 1 for the references listed in the reference section of all relevant and review articles identified by the electronic database search. To identify ongoing research, a search was completed for the inventory of Canadian Agri-Food Research [25] and TEKTRAN, which is a database that contains recent articles of published or soon-to-be published research results of the Agricultural Research Service [26] – the U.S. Department of Agriculture’s chief scientific research agency – but no related papers were found.

Population
Chicken meat	Laboratory media
(poultry or chicken)	(laboratory)
(broiler or meat)	(laboratory media or matrix)
(raw chicken)	(nutrient agar)
(chicken or broiler or flocks)	(BHI or brain–heart infusion)
(chicken production or processing)	(agar)
(chicks or chicken or parts or flocks)	(agar medium)
(chicken legs or wings or breasts or thighs or liver)	(TSA or TSB)
(minced or ground meat)	(nutrient or laboratory)
(skin-on or skin-off or chicken)
Refrigerated storage
(refrigeration)
(refrigeration time or temperature)
(chilling or chilled storage)
(refrigerated poultry or foods)
(refrigerated chicken or poultry)
(growth temperature)
(chilled foods)
(refrigeration or cooling)
(storage time or temperature)
(refrigerated storage)
(growth or survival)
Thermal inactivation
(cooking time or temperature)
(roasting or heating)
(D-values or Z-values)
(thermal treatment)
(thermal lethality)
(thermal or cooking inactivation)
(thermal inactivation or thermal resistance)
(cook or cooking)
Outcome
(enteric illness)
(foodborne disease or poisoning)
(disease or illness or risk)
(bacterial or pathogenic count or enumeration)
(infection or illness)
(Salmonella or concentration)
(bacteria or Salmonella or zoonoses)
(bacterial or bacteria load or level or counts)
(mesophilic counts or cfu)

a

Search terms were combined within each category using “OR” and between different categories using “AND”.

Table 1

Search terms used to identify potentially relevant literature to address quantitative effect of refrigerated storage and thermal inactivation temperatures on the enumeration of Salmonella on raw chicken meat and laboratory media.^a

2.2. Title and abstract screening

Abstracts were screened by one reviewer (H. Smadi) for relevance to the study objectives. An abstract was considered relevant if it described primary (original) research and evaluated Salmonella in fresh chicken meat or laboratory media, and Salmonella concentration during refrigerated storage or at different thermal inactivation temperatures (CFU/g or D-values) rather than prevalence was the focus of the research.

2.3. Full text screening

Full articles were obtained for all relevant abstracts and underwent full text screening by one reviewer (H. Smadi). Inclusion criteria for Salmonella growth studies were: (i) the study included one or more trials evaluating growth at temperatures ⩽10 °C; (ii) the study was conducted under normal storage conditions (e.g., air); and (iii) the study provided growth curves or data to calculate a growth curve (e.g., initial inoculation level, time or testing intervals and their corresponding log CFU/g) and storage temperatures were provided to model lag, exponential and stationary phases of the growth curves. The whole data set was used to model Salmonella density (Log₁₀ CFU/g) at the studied temperatures. If only generation time (GT) or growth rate (GR) were given to represent the exponential part of the growth curve, then the article was excluded. For inactivation studies, the inclusion criteria were: (i) initial inoculation level; (ii) thermal inactivation temperatures; and (iii) their corresponding D-values.

2.4. Quality assessment

Full papers were obtained for all relevant articles and underwent quality assessment. Standardized quality assessment forms were created for experimental studies. Only challenge trials, where Salmonella was inoculated directly onto the chicken meat and experiments took place in a controlled laboratory setting, and field trials with a natural exposure to Salmonella were found. No observational studies were identified. The following criteria were used to assess the study quality: description of randomization, blinding of the outcome assessor, numbers lost to follow-up, and measures of outcome variability or sufficient data to calculate one (standard deviation, standard error, confidence interval or P-value for post hoc calculation).

2.5. Data extraction

For all studies that passed the relevant screening stage, data extraction was conducted. When multiple trials were included within the same publication, data were extracted separately for each trial. After reviewing the articles, it was found that the longest testing intervals for chicken were 15 days, whereas it was 50 days for laboratory media. Therefore, in this analysis, only data up to 15 days were included for laboratory media to allow comparisons to chicken data. Repeated measures designs were used for all laboratory media studies, whereas none of the chicken data studies used repeated measures, i.e., different chicken pieces were used to enumerate Salmonella levels at each of the time periods evaluated.

For studies evaluating thermal inactivation of Salmonella, data on whether or not Salmonella Senftenberg was part of the inoculum were included because this serotype is more heat resistant than other Salmonella serotypes commonly used in thermal tests [27,28]. Two studies [27,28] reported the inoculation level as a range of Log₁₀ 7–8 and one study [29] reported it as a range of Log₁₀ 7–7.5. As a conservative assumption, in the statistical analysis, it was assumed that the inoculation level for these studies was Log₁₀ 7 CFU/g as typically this is the spoilage level for bacterial pathogens in raw poultry [30]. True replicates were used at each of the testing temperatures for both chicken and laboratory data.

2.6. Statistical methods

Weighted least squares regression analyses were performed independently for Log₁₀ Salmonella CFU and Log₁₀ D-values as the dependent variables for growth and inactivation, respectively. For growth, the outcome modeled was the change in Salmonella density (Log₁₀ CFU/g). Refrigerated storage temperatures (in Celsius), time (in days), and media type (chicken versus laboratory media) were modeled as the independent variables (fixed effects). Inoculation level was analyzed as part of the lag phase at time zero. Studies, trials within studies, and repeated measurements at a given temperature within an experiment were modeled as random effects. The covariance structure type was variance component. For repeated measures, different covariance structures were tried; however, the regression models did not converge. As a result, repeated measures were treated as a random effect to account partially for the correlation between measurements over time. For inactivation, Log₁₀ D-values were the dependent variable of interest. Temperature (in Celsius), media type (chicken versus laboratory media), inoculation levels (Log₁₀ CFU/g), and use of Salmonella Senftenberg (yes or no) were modeled as fixed effects and the study was included as a random effect.

For growth and inactivation, the assumption of heterogeneity between studies (or homogeneity within the same study) implied by the use of random effect is plausible due to different laboratory settings among different studies. To check the degree of similarity of growth of Salmonella between studies, the intra-class correlation coefficient ρ was computed, which estimates the proportion of the total variance that is due to the heterogeneity among studies [31].

Initially, all second and third order interaction terms and quadratic polynomial functions were evaluated for significance. The significance level used was 0.05. If the coefficient of a variable was not significant (P > 0.05), it was removed from the model using a backward regression. For the computation of the pooled effect estimates, each primary study was given a weight equal to the reciprocal of its sample size (SS) since the variance or data for post hoc calculation of the variance (standard error or standard deviation) were not provided in most studies. A normal distribution assumption was considered for the log of measured bacterial counts (and Log₁₀ D-values) for all treatment groups in all studies included in the meta-analysis. Normality tests of residuals, such as Shapiro–Wilk (W), Kolmogorove–Smirnov (D), Cramer–von Mises (W-Sq) and Anderson–Darling (A-Sq), and diagnostic plots of the residuals were used to test this assumption along with tests for skewness and kurtosis of the residuals.

3. Results

3.1. Identification and assessment of relevant literature

The search located 449, 395, 207, and 225 records for Salmonella growth in chicken, growth in laboratory media, thermal inactivation in chicken, and thermal inactivation in laboratory media, respectively. After relevance screening of titles and abstracts and removal of duplicated records, there were 18, 16, 6, and 4 records for each of these topic areas, respectively. Additional articles were excluded upon full text screening as follows. For growth in chicken, 10 articles were excluded because they reported prevalence only rather than concentration data and one [32] was excluded because it investigated growth in chicken à la king rather than raw chicken. For growth in laboratory media, four studies were excluded [33–36] because they reported prevalence rather than concentration data, and an additional four studies were excluded [10,22,37–38] because the temperatures evaluated were higher than the range of temperatures of interest in this review. Two additional studies were excluded: one study [39] gave only the generation time (GT) and the other [9] gave only the growth rate (GR) instead of full details of the growth curve. Therefore, there were 7, 6, 6, and 4 records for each of growth in chicken, growth in laboratory media, inactivation in chicken, and inactivation in laboratory media, respectively, included in the analysis (Tables 2–5).

Author	Study type (country)	Food type	Salmonella spp.	Enumeration method	Sample size/# time points tested	Temperature (inoculation level Log₁₀ CFU/g)
Pintar et al. [46]	Challenge (Canada)	Retail raw chicken meat (skinless, boneless chicken breast)	Typhimurium	USDA/FSIS MPN^a described in the Microbiology Laboratory Guidebook (sensitive to 0.3 MPN/g of sample)	14 (3)	4 °C (3.29)
CCFRA [41]	Challenge (U.K.)	Fresh boneless chicken thighs	Typhimurium, and Enteritidis	Selective media: XLD and MLCB	5 (6–9)	0, 4, 6, 8, 10 °C (3.79–4.54)
Baker et al. [40]	Challenge (U.S.)	Minced chicken meat	Typhimurium	Pour plate method	2 (6)	2, 7 °C (4)
	Field trial	Chicken parts (chicken breast muscle, and leg quarters with skin)	Enteritidis	Plate count agar (DIFCO)	2 (6)	2, 7 °C (N/A)
Gray et al. [45]	Challenge (U.S.)	Fresh chicken thighs	Enteritidis	Brilliant green agar (DIFCO)	2 (4)	10 °C (3.3)
Cunningham [44]	Challenge (U.S.)	Fresh broiler drumsticks	Typhimurium NRRL B-411	Salmonella-Shigella agar (SS agar)	10 (4)	10 °C (4.3)
To and Robach [43]	Challenge (U.S.)	Fresh whole broilers	Enteritidis ATCC 13076, Heidelberg ATCC 8326, Infantis 2-13, and Typhimurium ATCC 13311	Bismuth Sulfite Agar (DIFCO)	3 (5–6)	3 °C (1.9–2)
Robach and Ivey [42]	Challenge (U.S.)	Chicken breasts	Typhimurium 13311, Heidelberg 8326, and Montevideo 8387	Trypticase soy broth (DIFCO)	6 (3–4)	10 °C (1.5–3.6)

a

Most probable number (microbiological testing method).

Table 2

Summary of primary studies evaluating Salmonella growth at refrigerated storage temperatures (chicken meat).

Author	Study type (country)	Media type	Salmonella spp.	Enumeration method	Sample size/# time points tested	Temperature (inoculation level Log₁₀ CFU/g)	pH/NaCl
Alcock [48]	Challenge (U.K.)	Broth medium	Anatum, Montevideo, Napoli, Panama, Saint-paul, Stanley, Agona, Bredeney, Enteritidis, Hadar, Infantis, Senftenberg, Typhimurium, and Virchow	Nutrient agar plate	2 (9)	8 °C (4.5)	6.4/NR^a
Baker et al. [40]	Challenge (U.S.)	Trypticase soy broth	Typhimurium	Nutrient broth (DIFCO)	2 (6)	2, 7 °C (4.2)	NR/NR
Alcock [63]	Challenge (U.K.)	Broth medium	Agona, Bredeney, Enteritidis, Hadar, Infantis, Senftenberg, Typhimurium, and Virchow	Nutrient agar plate	2 (5–7)	6.2, 6.7, 7.6, 8.6 °C (4)	NR/NR
Elliott and Gray [49]	Challenge (Norway)	Trypticase soy agar	Enteritidis	TSA plates	2 (6)	10 °C (7)	6/NR
Matches and Liston [64]	Challenge (U.S.)	Nutrient broth	Heidelberg, Typhimurium, and Derby	Samples at 8 °C by TSA plates. At 12, 22, 37, and 41 °C by Bausch and Lomb Spectronic 20 spectrophotometer	2 (6)	8 °C (5.5)	NR/NR
Matches and Liston [65]	Challenge (U.S.)	Trypticase soy broth and agar	Derby, Heidelberg, Typhimurium, Aertrycke, Montevideo, Newport, and Thompson	TSA plates	2 (4)	5.1, 5.9, 6.7, 7.5, 8.3 °C (6.7)	NR/NR

a

NR: not reported.

Table 3

Summary of primary studies evaluating Salmonella growth at refrigerated storage temperatures (laboratory media).

Author	Study type (country)	Food type	Salmonella spp.	Enumeration method	Sample size per testing temperature	Temperature (inoculation level Log₁₀ CFU/g)
Murphy et al. [28]	Challenge (U.S.)	Chicken patties, and chicken tenders	Senftenberg, Typhimurium, Heidelberg, Mission, Montevideo and California	Food and Drug Administration procedures	3	55, 57.5, 60, 62.5, 65, 67.5, 70 °C (7–8)
Juneja et al. [50]	Challenge (U.S.)	Ground chicken.	Kentucky, Heidelberg, Hadar and Thompson	TSA surface plating	2	58, 60, 62.5, 65 °C (8)
Juneja et al. [52]	Challenge (U.S.)	Ground chicken	Thompson, Enteritidis, Typhimurium, Hadar, Copenhagen, Montevideo and Heidelberg	TSA surface plating	2	58, 60, 62.5, 65 °C (8)
Mazzotta [51]	Challenge (U.S.)	Ground chicken breast meat	Typhimurium, Enteritidis, Montevideo, Mbandaka, Heidelberg and Thompson	TSA surface plating	3	56, 60, 62, 63 °C (7)
Murphy et al. [29]	Challenge (U.S.)	Ground chicken breast meat	Senftenberg, Typhimurium, Heidelberg, Mission, Montevideo and California	Food and Drug Administration procedures	3	55, 57.5, 60, 62.5, 65, 67.5, 70 °C (7–7.5)
Murphy et al. [27]	Challenge (U.S.)	Ground chicken breast meat	Senftenberg, Typhimurium, Heidelberg, Mission, Montevideo and California	Food and Drug Administration procedures	3	67.5, 70 °C (7–8)

Table 4

Summary of primary studies evaluating Salmonella inactivation at thermal treatment temperatures (chicken meat).

Author	Study type (country)	Media type	Salmonella spp.	Enumeration method	Sample size per testing temperature	Temperature (inoculation level Log₁₀ CFU/g)
Juneja et al. [50]	Challenge (U.S.)	Broth medium	Thompson, Enteritidis, Typhimurium, Hadar, Copenhagen, Montevideo and Heidelberg	TSA surface plating	2	55, 58, 60, 62 °C (8)
Murphy et al. [29]	Challenge (U.S.)	Liquid medium (0.1% peptone-agar solution)	Senftenberg, Typhimurium, Heidelberg, Mission, Montevideo and California	Food and Drug Administration procedures	3	55, 57.5, 60, 62.5, 65, 67.5, 70 °C (7–7.5)
Murphy et al. [27]	Challenge (U.S.)	Liquid medium (0.1% peptone-agar solution)	Senftenberg, Typhimurium, Heidelberg, Mission, Montevideo and California	Food and Drug Administration procedures	3	67.5, 70 °C (7–8)
Xavier and Ingham [66]	Challenge (Canada)	Casein soymeal peptone-yeast extract broth medium	Enteritidis (ATCC4931)	Nutrient agar (BDH)	2	52, 54, 56, 58 °C (7)

Table 5

Summary of primary studies evaluating Salmonella inactivation at thermal treatment temperatures (laboratory media).

3.1.1. Refrigerated storage

Tables 2 and 3 summarize characteristics of the studies that met the inclusion criteria for Salmonella growth on chicken and laboratory media at refrigerated storage temperatures, respectively. For chicken, all articles were challenge trials, except one [40], which had both a challenge trial and a field trial within the same publication. Combining results from different study designs was not performed, as an invalid effect estimate may arise due to the methodological heterogeneity of different designs [24]. As a result, the field trial was excluded from the meta-analysis. Among the seven studies included, three trials had more than one challenge experiment within the same publication [41–43]. CCFRA [41] tested Salmonella growth in chicken using two different enumeration media and had two sets of five replicates for each medium. Results from all replicates in different media were included separately in the analysis. To and Robach [43] tested Salmonella growth at two poultry processing plants. Results from both plants were included separately in the analysis. Robach and Ivey [42] reported two trials with two different levels of Salmonella inoculation. Both trials were included in our analysis. For studies investigating the effect of different interventions, such as carbon dioxide and potassium sorbate on Salmonella growth [40,42–45], only control growth curves (e.g., air) were included in this analysis. All studies reported the outcome as a continuous outcome, e.g., at each storage temperature; testing intervals and their corresponding Log₁₀ CFU/g were reported. Only one study reported growth of Salmonella as most probable number (MPN)/g [46] which was converted into CFU/g for analysis [47].

Use of randomization was explicitly stated in only one trial [46]. Blinding of the person assessing the outcome and loss to follow-up were not reported in any trial. Standard deviation or variability measures for post hoc calculation were reported only in one study [41]. Therefore, evaluating the impact of these factors was not possible. For laboratory media, only a few studies controlled for the potential confounding effect of pH value [48,49], and none of the studies reported NaCl level, which is a potentially confounding variable, when examining the relationship between temperature and growth of Salmonella. Enumeration methods used to count Salmonella were all standard cultural methods.

3.1.2. Thermal inactivation

Tables 4 and 5 summarize characteristics of the studies that met the inclusion criteria for thermal inactivation studies in chicken and laboratory media, respectively. All studies reported the initial inoculation level, testing temperatures and their corresponding D-values. All were challenge trials and estimated D-values over a range of temperatures.

One study explicitly stated random allocation to treatment [50]. Blinding of the person assessing the outcome and loss to follow-up were not reported in any of the studies. Standard deviation or variability measures for post hoc calculation were provided in three of six trials [27,50–51]. Three studies [27,50–51] reported standard deviation within each treatment group and three studies [28–29,52] did not report the SD or data needed for post hoc calculation. One study that did report the SD used S. Senftenberg [27]; the other two studies that reported the SD did not use S. Senftenberg [50,51]. Among the studies that did not report the SD [28–29,52,], only one did not use S. Senftenberg [52].

3.2. Meta-analysis equations

3.2.1. Refrigerated storage

Significant predictors of Salmonella growth (Log₁₀ CFU) in chicken meat versus laboratory media are shown in Table 6. The intra-class correlation coefficient, ρ, calculated as the ratio of the estimate of the study random effect divided by the sum of the estimates of all variance components was 0.51. This indicated that there was a similarity in growth of Salmonella within the same study and therefore the study should be included as a random effect. The residual in this table was significant (<0.001), meaning that growth varied significantly within studies and trials even after controlling for the other effects in the model.

Cov parameter	Estimate	Standard error	Z value	Pr > Z	Alpha	Lower CI^a	Upper CI
Estimates of covariance parameters
Study(Media)	1.03	0.92	1.13	0.13	0.05	0.31	20.67
Trial(Study × Media)	0.11	0.095	1.13	0.13	0.05	0.032	2.18
Repeated(Study × Media × Temperature × Trial)	0.66	0.11	6.19	<0.0001	0.05	0.49	0.93
Residual	0.23	0.045	5.05	<0.0001	0.05	0.16	0.35

Effect	Num DF	Den DF	F value	Pr > F
Estimates of fixed effects
Temperature	12	87	0.89	0.56
Time	1	50	65.02	<0.0001
Media	1	4	0.12	0.74
Time × Temperature	12	50	25.67	<0.0001
Media × Temperature	2	87	1.27	0.29
Time × Media	1	50	0.48	0.49
Time × Media × Temperature	2	50	6.39	0.0034
Time × Time	1	50	8.74	0.0047

a

CI: confidence interval.

Table 6

Results of meta-analysis equation for growth of Salmonella in chicken meat and laboratory media over 15 days at temperatures ⩽10 °C.

The significant three-way interaction term Time × Media × Temperature indicated that there was a significant difference in the growth of Salmonella on chicken versus that on laboratory media under the same storage time and temperature. The ratio of the variance estimate between the two media (chicken divided by laboratory) was 6.79 (data not shown). This means that Salmonella growth on chicken varied 6.79 times more than in laboratory media. Thus Salmonella growth in the two media differed in the average value, and there was more variation on the pattern of growth across different testing intervals between the two media types.

Therefore, chicken data were analyzed alone to estimate the growth equations at different temperatures. Table 7 shows the parameter estimates for the random effect covariance, fixed effects, and solutions for fixed effects for chicken data. There was a lack of a consistent pattern in growth with the increase in temperature. Growth at temperatures 4 °C and 7 °C was significantly different than growth at other temperatures (P = 0.005), otherwise there were no differences among the remaining temperatures. Therefore, it is more appropriate to model temperatures separately rather than combining them in a single estimate, and the results are presented as such.

Cov parameter	Estimate	Standard error	Z value	Pr > Z	Alpha	Lower CI	Upper CI
Estimates of covariance parameters
Study	0.87	0.801	1.09	0.14	0.05	0.25	21.23
Trial (Study)	0.096	0.094	1.02	0.15	0.05	0.026	3.44
Residual	3.45	0.53	6.54	<0.0001	0.05	2.61	4.77

Effect	Num DF	Den DF	F value	Pr > F
Estimates of fixed effects
Temperature	5	84	1.65	0.15
Time	1	84	29.2	<0.0001
Time × Temperature	5	84	2.17	0.065
Time × Time	1	84	15.82	0.0001

Effect	Temperature	Estimate	Standard error	DF	t value	Pr > l t l
Solution for fixed effects
Temperature	2	4.02	1.37	84	2.94	0.004
Temperature	3	1.43	1.11	84	1.28	0.203
Temperature	4	3.21	1.09	84	2.94	0.004
Temperature	7	4.27	1.37	84	3.12	0.003
Temperature	8	4.91	0.62	84	7.93	<0.0001
Temperature	10	4.63	0.55	84	8.46	<0.0001
Time × Temperature	2	0.45	0.32	84	1.43	0.16
Time × Temperature	3	0.53	0.12	84	4.6	<0.0001
Time × Temperature	4	0.22	0.10	84	2.18	0.03
Time × Temperature	7	1.005	0.32	84	3.18	0.002
Time × Temperature	8	0.47	0.099	84	4.72	<0.0001
Time × Temperature	10	0.45	0.071	84	6.36	<0.0001
Time × Time		−0.027	0.007	84	−3.98	0.0001

Table 7

Results of meta-analysis equation for growth of Salmonella in chicken meat over 15 days at temperatures ⩽10 °C.

Statistical normality tests were all <0.05 indicating that the residual data were not normally distributed (e.g., P-values for W, D, W-Sq, and A-Sq were <0.0001, <0.01, <0.005, and <0.005, respectively). Skewness and Kurtosis were −1.51 and 5.70, respectively, indicating that the residuals were skewed to the left with a peaked curve. To adjust for this non-normality in the residuals distribution log–log CFU/g and square root of CFU/g transformations were evaluated. However, none of these transformations resulted in normally distributed residuals and were therefore not used in the final model.

3.2.2. Thermal inactivation

A comparison of the parameter estimates for temperature at different inoculation levels (7 or 8), media type (Chicken = C, Laboratory = L), whether Salmonella Senftenberg was part of the serotypes mix (Yes = Y, or No = N), and significance of second and third interaction terms and their effect on Log₁₀ D-values during heat treatment is summarized in Table 8.

Cov parameter	Estimate	Standard error	Z value	Pr > Z	Alpha	Lower CI^a	Upper CI
Estimates of covariance parameters
Study(Media × Senftenberg)	0.008	0.007	1.18	0.12	0.05	0.003	0.13
Residuals	0.019	0.003	6.26	<0.0001	0.05	0.014	0.03

Effect	Num DF	Den DF	F value	Pr > F
Estimates of fixed effects
Media	1	5	8.69	0.03
Senftenberg	1	5	18.37	0.008
Inoculation	1	79	5.38	0.02
Temperature	1	79	15.97	0.0001
Media × Senftenberg	1	5	3.15	0.14
Temperature × Media	1	79	13.44	0.0004
Temperature × Senftenberg	1	79	18.77	<0.0001
Temperature × Media × Senftenberg	1	79	4.76	0.03
Temperature × Temperature	1	79	26.55	<0.0001
Temperature × Inoculation	1	79	8.81	0.004
Temperature × Temperature × Senftenberg	1	79	20.16	<0.0001

Effect	Media/Senftenberg^b	Inoculation	Estimate	Standard error	DF	t value	Pr > l t l
Solution for fixed effects
Inoculation		7	6.65	2.65	79	2.51	0.01
Inoculation		8	4.42	2.82	79	1.57	0.12
Media × Senftenberg	C/N		−34.49	8.06	5	−4.28	0.008
Media × Senftenberg	C/Y		−0.64	0.45	5	−1.42	0.21
Media × Senftenberg	L/N		−31.92	7.31	5	−4.37	0.007
Media × Senftenberg	L/Y		0
Temperature × Media × Senftenberg	C/N		1.17	0.27	79	4.38	<0.0001
Temperature × Media × Senftenberg	C/Y		0.006	0.09	79	0.07	0.95
Temperature × Media × Senftenberg	L/N		1.11	0.25	79	4.41	<0.0001
Temperature × Media × Senftenberg	L/Y		−0.008	0.086	79	−0.09	0.93
Temperature × Inoculation		7	−0.05	0.016	79	−2.97	0.004
Temperature × Inoculation		8	0
Temperature × Temperature × Senftenberg	N		−0.01	0.002	79	−5.05	<0.0001
Temperature × Temperature × Senftenberg	Y		−0.00075	0.0007	79	−1.12	0.27

a

CI: confidence interval.

b

C: Chicken media, L: Laboratory media, Y: (Yes) multiple Salmonella serotypes including Salmonella Senftenberg, N: (No) multiple Salmonella serotypes without Salmonella Senftenberg.

Table 8

Results of meta-analysis equation for thermal inactivation of Salmonella in chicken meat and laboratory media at temperatures ranging from 55 to 70 °C.

For thermal inactivation equations, at different combinations, the intercepts and slopes, respectively were estimated to be: (7, C, Y) 6.016 and −0.043 (7, C, N), −27.84 and 1.12 (7, L, Y). 6.65 and −0.056 (7, L, N), −25.27 and 1.066 (8, C, Y), 3.78 and 0.0057 (8, C, N), −30.074 and 1.168 (8, L, Y), 4.418 and −0.0077 (8, L, N), −27.51 and 1.115. From this data, D-values were consistently higher in chicken meat than in laboratory media, when S. Senftenberg was part of the inoculation and when the inoculation level was 8 in comparison to 7. Also, as the temperature increased, D-values decreased. Normality tests were all >0.05, meaning that the residuals from the inactivation model did not show any significant normality (e.g., P-values for W, D, W-Sq, and A-Sq were 0.28, >0.15, >0.25, and >0.25, respectively). Skewness and Kurtosis were −0.27 and −0.57 indicating that the data were a bit skewed to the left with small flatness in the curve.

4. Discussion

This review found a significant difference between growth/inactivation in chicken meat versus that in laboratory media. As a result, the use of laboratory media as an alternative to chicken meat in QRA modeling may not be appropriate. Meta-analysis equations for Salmonella growth and inactivation in chicken meat were developed in this study and could be used to support future risk assessment modeling to estimate growth and inactivation of Salmonella at different temperatures. For chicken, there was no consistent pattern in growth of Salmonella as a function of temperature. No specific explanation could be found considering that all the studied temperatures were at the lower scale of Salmonella growth temperatures (e.g., ⩽10 °C). However, there was evidence of growth of Salmonella at temperatures of less than 10 °C. Ignoring growth in this temperature range may therefore underestimate the total number of Salmonella predicted in chicken that can cause the illness.

In this review, growth data (starting from zero days until the maximum of 15 days) were modeled rather than growth rate, which is commonly used to model bacterial growth [53]. This can have several advantages: (i) this approach avoids the use of subjective measures to decide on cut-points between the end and start of different growth phases, especially when growth does not follow the traditional sigmoidal shape and (ii) it includes all data points available to model different growth phases (including the lag phase), which makes biological sense as bacteria require time to adapt when moved from one environment to another, while using growth rate will overestimate bacterial numbers predicted in the lag phase.

As the temperature increased, D-values for Salmonella inactivation decreased. For thermal inactivation modeling in chicken meat, it is recommended to use the thermal inactivation equations presented with Salmonella Senftenberg included in the inoculum, as the high thermal resistance of this serotype will provide a conservative assumption of killing other serotypes contaminating the chicken meat. However, the three studies pooled for Salmonella Senftenberg were from the same author. This might enhance consistency of the testing environment, such as the laboratory setting, source of chicken pieces tested, reliability of tools used to measure the outcome, and source of bacteria (age of culture) used to contaminate the chicken.

4.1. Methodological issues/sources of heterogeneity

The current review may be subject to selection bias as the focus was only on English language articles [54]. However, an attempt was made to reduce selection bias in the identification of primary research studies by: (i) searching six databases that are among the most commonly used in the food safety area; (ii) expanding the publication period of the included trials; and (iii) contacting experts in the field to identify unpublished work. Furthermore, most of the studies found were challenge studies in which case it would not be expected to find different results in studies from different geographic areas. However, challenge trials do not provide evidence of a high quality for real world application as do natural disease outcomes [55], and hence having more studies with natural exposure to Salmonella to address these types of questions is an area to consider when designing future studies.

Methodological concerns were identified for several studies in the quality assessment stage. Randomization, blinding and loss-to-follow-up were generally not reported, raising the possibility of selection bias (at the chicken parts level). It is possible that random allocation was performed, but not explicitly reported. Blinding may not have been used because objective laboratory techniques were used to determine the outcome, so the laboratory technicians’ knowledge of the treatment temperatures would not likely affect the measured outcome, and loss-to-follow-up (due, for example, to mishandling practices or spoilage) may not have been reported since this type of research lasts from a few days to a few weeks at most due to the short shelf life of the tested products. Nonetheless, the use of guidelines in food safety research, such as the CONSORT and REFLECT statements, may assist in ensuring complete reporting of essential design features for RCTs [56–59] and similarly for challenge trials.

Clinical heterogeneity might exist in the combined studies due to variability in the chicken characteristics. For example, in refrigerated storage studies, different types of chicken meat were investigated, such as skinless, boneless chicken breast, chicken thighs, and chicken muscles. Similarly, the thermal inactivation studies included chicken patties, chicken tenders, ground chicken, and ground chicken breast. Different chicken parts may vary in their pH values [41]. Thigh chicken meat, for example, has a higher pH value (6.4–6.7) than breast chicken meat (5.8) [41] which makes the former closer to the optimum pH required for Salmonella to grow which lies between 6.5 and 7.5 [18]. Subgroup analysis to compare growth/inactivation in different chicken types was not performed due to the limited number of studies.

Sub-group analysis to investigate the impact of potential confounders and effect modifiers was not performed as the number of trials available was insufficient. Potential confounding variables include pH and water activity level, and an example of an effect modifier is the percentage of fat level in the tested product. The higher the percentage fat, the higher the time (D-value) needed to heat the product to a certain temperature [60,61]; and the higher percentage fat, the higher the protection for bacterial cells against heat [62]. Other examples might be the history and age of the Salmonella mixture inoculated and level of nutrients available for Salmonella to grow. Control for such factors, by measuring the composition of the tested products and deriving separate equations to apply to different levels of the effect variables, minimizes the possibility of invalid effect estimates.

5. Conclusion

The current meta-analysis approach provided a structured method for finding and pooling data to increase precision of estimates for Salmonella growth and inactivation at different temperatures. A significant difference was found between Salmonella growth/inactivation on chicken meat versus laboratory media. The growth and inactivation meta-analysis equations detailed in this review should be used in QRA to model growth and inactivation of Salmonella in chicken meat. Parameter estimates for growth of Salmonella in chicken meat at temperatures ⩽10 °C and inactivation at temperatures between 55 °C and 70 °C were provided and should be used when modeling Salmonella growth and inactivation in chicken meat. Validation of growth/inactivation equations created in this review against independent data is an area to be considered for future research, keeping in mind the methodological recommendations made in this paper to enhance the quality of reported data in the food safety area.

Conflict of interest

There are no competing interests.

Acknowledgments

The authors thank William Sears for assistance with the statistical analysis. Funding for this project was received from the Laboratory for Foodborne Zoonoses, Public Health Agency of Canada, and the Canadian Institutes of Health Research (CIHR) Institute of Population and Public Health/Public Health Agency of Canada Applied Public Health Research Chair.

References

[1]Public Health Agency of Canada (PHAC), Notifiable diseases on-line, Salmonellosis, 2003. Available from: http://dsol-smed.phac-aspc.gc.ca/dsol-smed/ndis/list_e.html.

[2]SA Fernandes, ACR Ghilardi, TA Tavechio, AMO Machado, and ACC Pignatari, Phenotypic and molecular characterization of Salmonella eneritidis strains isolated in Sao Paulo, Brazil, Rev Inst Med Trop Sao Paulo, Vol. 45, No. 2, 2003, pp. 59-63.

[3]Health Canada, Salmonellosis-notifiable diseases on-line, 1998. Available from: http://dsol-smed.phac-aspc.gc.ca/dsol-smed/ndis/diseases/salm_e.html.

[4]A Hogue, P White, JG Petter, N Schlosser, R Gast, E Ebel, et al., Epidemiology and control of egg-associated with Salmonella enteritidis in the United States of America, Rev Sci Tech, Vol. 16, 1997, pp. 542-53.

[5]GC Mead, Microbiological quality of poultry meat: a review, Braz J Poult Sci, Vol. 6, No. 3, 2004, pp. 135-42.

[6]P Zhao, T Zhao, MP Doyle, JR Rubino, and J Meng, Development of a model for evaluation of microbial cross-contamination in the kitchen, J Food Prot, Vol. 61, No. 8, 1998, pp. 960-3.

[7]Codex Alimentarius Commission (CAC), Principles and guidelines for the conduct of a microbial risk assessment, FAO, Rome, 1999. CAC/GL-30.

[8]JM Broughall and C Brown, Hazard analysis applied to microbial growth in foods: development and application of three-dimensional models to predict bacterial growth, Food Microbiol, Vol. 1, 1984, pp. 13-22.

[9]AM Gibson, N Bratchell, and TA Roberts, Predicting microbial growth: growth responses of Salmonella in a laboratory medium as affected by pH, sodium chloride and storage temperature, Int J Food Microbiol, Vol. 6, 1988, pp. 155-78.

[10]DW Thayer, WS Muller, RL Buchanan, and JG Phillips, Effect of NaCl, pH, temperature, and atmosphere on growth of Salmonella typhimurium in glucose-mineral salts medium, Appl Environ Microbiol, Vol. 53, 1987, pp. 1311-5.

[11]TP Oscar, Response surface models for effects of temperature, pH, and previous growth pH on growth kinetics of Salmonella Typhimurium in brain-heart infusion broth, J Food Prot, Vol. 62, 1999, pp. 106-11.

[12]TP Oscar, Validation of a tertiary model for predicting variation of Salmonella typhimurium DT104 (ATCC 700408) growth from a low initial density on ground chicken breast meat with a competitive microflora, J Food Prot, Vol. 69, No. 9, 2006, pp. 2048-57.

[13]TP Oscar, Response surface models for effects of temperature and previous growth sodium chloride on growth kinetics of Salmonella Typhimurium on cooked chicken breast, J Food Prot, Vol. 62, No. 12, 1999, pp. 1470-4.

[14]TP Oscar, Response surface models for effects of temperature and previous temperature on lag time and specific growth rate of Salmonella Typhimurium on cooked ground chicken breast, J Food Prot, Vol. 62, No. 10, 1999, pp. 1111-4.

[15]TP Oscar, Variation of lag time and specific growth rate among 11 strains of Salmonella inoculated onto sterile ground chicken breast burgers and incubated at 25 °C, J Food Saf, Vol. 20, 2000, pp. 225-36.

[16]TP Oscar, Development and validation of a tertiary simulation model for predicting the potential growth of Salmonella typhimurium on cooked chicken, Int J Food Microbiol, Vol. 76, 2002, pp. 177-90.

[17]TP Oscar, Predictive models for growth of Salmonella typhimurium DT104 from low and high initial density on ground chicken with a natural microflora, Food Microbiol, Vol. 24, 2007, pp. 640-51.

[18]JY D’Aoust, Salmonella, MP Doyle (editor), Foodborne bacterial pathogens, Marcel Dekker, Inc., New York, 1989, pp. 327-445.

[19]D Roberts, Sources of infection: food, WM Waites and JP Arbuthnott (editors), Foodborne illness, Edward Arnold, London, 1991, pp. 31-7.

[20]ECD Todd, Factors that contributed to foodborne disease in Canada, 1973–1977, J Food Prot, Vol. 46, No. 8, 1983, pp. 737-47.

[21]BD Gessner and M Beller, Protective effect of conventional cooking versus use of microwave ovens in an outbreak of salmonellosis, Am J Epidemiol, Vol. 139, No. 9, 1994, pp. 903-9.

[22]VK Juneja, HM Marks, and L Huang, Growth and heat resistance kinetic variation among various isolates of Salmonella and its application to risk assessment, Risk Anal, Vol. 23, No. 1, 2003, pp. 199-213.

[23]ME Doyle and AS Mazzotta, Review of studies on the thermal resistance of salmonellae, J Food Prot, Vol. 63, 2000, pp. 779-95.

[24]KS Khan, R Kunz, J Kleijnen, and G Antes, Systematic reviews to support evidence-based medicine, How to review and apply findings of healthcare research, The Royal Society of Medicine Press Limited, 2003.

[25]Inventory of Canadian Agri-Food Research (ICAR). no date. Available from: http://publications.gc.ca/control/publicationInformation?searchAction=2&publicationId=275013&lang=English&s=new.

[26]Agricultural Research Service (ARS), the U.S. Department of Agriculture’s chief scientific research agency. no date. Available from http://www.ars.usda.gov/services/tektran.htm.

[27]RY Murphy, BP Marks, ER Johnson, and MG Johnson, Inactivation of Salmonella and Listeria in ground chicken breast meat during thermal processing, J Food Prot, Vol. 62, No. 9, 1999, pp. 980-5.

[28]RY Murphy, LK Duncan, ER Johnson, MD Davis, and JN Smith, Thermal inactivation D- and z-values of Salmonella serotypes and Listeria innocua in chicken patties, chicken tenders, franks, beef patties, and blended beef and turkey patties, J Food Prot, Vol. 65, No. 1, 2002, pp. 53-60.

[29]RY Murphy, BP Marks, ER Johnson, and MG Johnson, Thermal inactivation kinetics of Salmonella and Listeria in ground chicken breast meat and liquid medium, J Food Sci, Vol. 65, No. 4, 2000, pp. 706-10.

[30]MB Mielnik, RH Dainty, F Lundby, and J Mielnik, The effect of evaporative air chilling and storage temperature on quality and shelflife of fresh chicken carcasses, Poult Sci, Vol. 78, 1990, pp. 1065-73.

[31]DS Moskowitz and SL Hershberger, Modeling intra-individual variability with repeated measures data: methods and applications, Lawrence Erlbaum Associates, Publishers, London, 2002.

[32]R Angelotti, MJ Foter, and KH Lewis, Time-temperature effects on salmonellae and staphylococci in foods. I. Behaviour in refrigerated foods, Am J Public Health, Vol. 51, No. 1, 1961, pp. 76-83.

[33]KC Chung and JM Goepfert, Growth of Salmonella at low pH, J Food Sci, Vol. 35, 1970, pp. 326-8.

[34]AM Gibson and TA Roberts, The effect of pH, water activity, sodium nitrite and storage temperature on the growth of enteropathogenic Escherichia coli and salmonellae in a laboratory medium, Int J Food Microbiol, Vol. 3, 1986, pp. 183-94.

[35]DAA Mossell, M Jansma, and J de Waart, Growth potential of 114 strains of epidemiologically most common Salmonellae and arizonae between 3 and 17 °C, TA Roberts, G Hobbs, JHB Chritian, and N Skorgaard (editors), Psychrotrophic microorganisms in spoilage and pathogenicity, Academic Press, New York, NY, 1981, pp. 29-37.

[36]JR Matches and J Liston, Effect of pH on low temperature growth of Salmonella, J Milk Food Technol, Vol. 35, No. 1, 1972, pp. 49-52.

[37]KY Li and JA Torres, Water activity relationships for selected mesophiles and psychrotrophs at refrigeration temperature, J Food Prot, Vol. 56, No. 7, 1993, pp. 612-5.

[38]LE Phillips, TJ Humphrey, and HM Lappin-Scott, Chilling invokes different morphologies in two Salmonella enteritidis PT4 strains, J Appl Microbiol, Vol. 84, 1998, pp. 820-6.

[39]K Fehlhaber and G Kruger, The study of Salmonella enteritidis growth kinetics using rapid automated bacterial impedance technique, J Appl Microbiol, Vol. 84, 1998, pp. 945-9.

[40]RC Baker, RA Qureshi, and JH Hotchkiss, Effect of an elevated level of carbon dioxide containing atmosphere on the growth of spoilage and pathogenic bacteria at 2, 7, and 13 °C, Poult Sci, Vol. 65, 1986, pp. 729-37.

[41]Campden & Chorleywood Food Research Association Group (CCFRA), Evaluating the growth and survival of Salmonella on chicken at chilled storage temperatures. Project number B12004, Agency contact for further information, 2003. Available at kathryn.callaghan@foodstandards.gsi.gov.uk.

[42]MC Robach and FJ Ivey, Antimicrobial efficacy of a potassium sorbate dip on freshly processed poultry, J Food Prot, Vol. 41, No. 4, 1978, pp. 284-8.

[43]EC To and MC Robach, Potassium sorbate dip as a method of extending shelf life and inhibiting the growth of Salmonella and Staphylococcus on fresh, whole broilers, Poult Sci, Vol. 59, 1980, pp. 726-30.

[44]FE Cunningham, Microbiology of poultry parts dipped in potassium sorbate, Poult Sci, Vol. 60, 1981, pp. 969-71.

[45]RJH Gray, PH Elliott, and RI Tomlins, Control of two major pathogens on fresh poultry using a combination potassium sorbate/carbon dioxide packaging treatment, J Food Sci, Vol. 49, 1984, pp. 142-6.

[46]K Pintar, A Cook, F Pollari, A Ravel, S Lee, and JA Odumeru, Quantitative effect of refrigerated storage time on the enumeration of Campylobacter, Listeria, and Salmonella on artificially inoculated raw chicken meat, J Food Prot, Vol. 70, No. 3, 2007, pp. 739-43.

[47]JA McCarthy, HA Thomas, and JE Delaney, Evaluation of the reliability of coliform density tests, Am J Public Health Nations Health, Vol. 48, No. 12, 1958, pp. 1628-35.

[48]SJ Alcock, Growth characteristics of food poisoning organisms at sub-optimal temperatures, Campden Food Preservation Research Association, 1987. Technical, Memorandum No. 440

[49]PH Elliott and RJH Gray, Salmonella sensitivity in a sorbate/modified atmosphere combination system, J Food Prot, Vol. 44, No. 12, 1981, pp. 903-8.

[50]VK Juneja, BS Eblen, and GM Ransom, Thermal inactivation of Salmonella spp. in chicken broth, beef, pork, turkey, and chicken: determination of D- and Z-values, J Food Sci, Vol. 66, No. 1, 2001, pp. 146-52.

[51]AS Mazzotta, D- and z-values of Salmonella in ground chicken breast meat, J Food Saf, Vol. 20, 2000, pp. 217-23.

[52]VK Juneja, BS Eblen, and HM Marks, Modeling non-linear survival curves to calculate thermal inactivation of Salmonella in poultry of different fat levels, Int J Food Microbiol, Vol. 70, 2001, pp. 37-51.

[53]SJC Van Gerwen and MH Zwietering, Growth and inactivation models to be used in quantitative risk assessments, J Food Prot, Vol. 61, No. 11, 1998, pp. 1541-9.

[54]Cochrane Handbook for Systematic Reviews of Interventions 4.2.6. Updated September 2006, The Cochrane Collaboration. Available from: http://www.cochrane.org/resources/handbook/.

[55]IR Dohoo and D Waltner-Toews, Interpreting clinical research part III. Observational studies and interpretation of results, Interpret Clin Res, Vol. 7, No. 10, 1985, pp. S605-12.

[56]C Begg, M Cho, S Eastwood, R Horton, D Moher, I Olkin, et al., Improving the quality of reporting of randomized controlled trial: the CONSORT statement, JAMA, Vol. 276, 1996, pp. 637-9.

[57]D Moher, CONSORT: an evolving tool to help improve the quality of reports of randomized controlled trials, JAMA, Vol. 279, 1998, pp. 1489-91.

[58]AM O’Connor, JM Sargeant, I Gardner, J Dickson, M Torrence, CE Dewey, et al., The REFLECT statement: methods and processes of creating reporting guidelines for randomized controlled trials for livestock and food safety, J Food Prot, Vol. 73, 2010, pp. 132-9.

[59]JM Sargeant, AM O’Connor, IA Gardner, JS Dickson, and ME Torrence, Consensus Meeting Participants: Dohoo IR, Lefebvre SL, Morley PS, Ramirez A, Snedeker K. The REFLECT Statement: reporting guidelines for randomized controlled trials in livestock and food safety: explanation and elaboration, J Food Prot, Vol. 73, 2010, pp. 579-603.

[60]MN Ahmed, DE Conner, and DL Huffman, Heat-resistance of Escherichia coli O157: H7 in meat and poultry as affected by product composition, J Food Sci, Vol. 60, 1995, pp. 606-10.

[61]AR Fain Jr., JE Line, AB Moran, LM Martin, RV Lechowich, JM Carosella, et al., Lethality of heat to Listeria monocytogenes Scott A: D-value and z-value determinations in ground beef and turkey, J Food Prot, Vol. 54, 1991, pp. 756-61.

[62]JM Jay, Modern food microbiology, 3rd ed., Van Nostrand Reinhold Co., New York, 1986.

[63]SJ Alcock, Survival and growth of Salmonellae at chill temperatures, in Proceedings of chilled food symposium (OCLC, 1985), pp. 1417-57.

[64]JR Matches and J Liston, Effects of incubation temperature on the salt tolerance of Salmonella, J Milk Food Technol, Vol. 35, No. 1, 1972, pp. 39-44.

[65]JR Matches and J Liston, Low temperature growth of Salmonella, J Food Sci, Vol. 33, 1968, pp. 641-5.

[66]IJ Xavier and SC Ingham, Increased D-values for Salmonella enteritidis following heat shock, J Food Prot, Vol. 60, No. 2, 1997, pp. 181-4.

Download article (PDF)

Next Article In Issue>

Journal: Journal of Epidemiology and Global Health
Volume-Issue: 2 - 4
Pages: 165 - 179
Publication Date: 2012/12/27
ISSN (Online): 2210-6014
ISSN (Print): 2210-6006
DOI: 10.1016/j.jegh.2012.12.001 How to use a DOI?
Open Access: This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)

Cite this article

ris enw bib

TY  - JOUR
AU  - Hanan Smadi
AU  - Jan M. Sargeant
AU  - Harry S. Shannon
AU  - Parminder Raina
PY  - 2012
DA  - 2012/12/27
TI  - Growth and inactivation of Salmonella at low refrigerated storage temperatures and thermal inactivation on raw chicken meat and laboratory media: Mixed effect meta-analysis
JO  - Journal of Epidemiology and Global Health
SP  - 165
EP  - 179
VL  - 2
IS  - 4
SN  - 2210-6014
UR  - https://doi.org/10.1016/j.jegh.2012.12.001
DO  - 10.1016/j.jegh.2012.12.001
ID  - Smadi2012
ER  -

download .riscopy to clipboard