o-

. H

b Irstea, Refrigeration Process Engineering Research Unit, 1 rue Pierre-Gilles de Gennes, 92761 Antony, FrancecADRIA Dveloppement, Technical Institute for Food Industry, ACTIA centre, Z.A. Creach Gwen, 29196 Quimper cedex, FrancedDivision of Mechatronics, Biostatistics and Sensors (Me

a r t i c l e i n f o

Article history:

according to their composition and properties. This is the case formeat and processed meat products which are ideal for the growthof spoilage and pathogenic bacteria. Pork meat and Ready-To-Eat(RTE) pork meals are the main type of meat consumed in Europe(Mataragas et al., 2008; Verbeke et al., 2010). Moreover, they wereidentied as one of the food products where the prevalence of path-

siness operatorsat the bac. In additioof RTE por

products make possible the growth of other types of bacteriarefrigerated temperatures (+2/+4 C): the lactic acid bacteriaA part of this LAB ora is responsible for spoilage and qualby inducing physico-chemical changes into the food products(Laursen et al., 2009; Hereu et al., 2012). The physico-chemicalmodications are often assessed through quality indicators likecolour, texture, water holding capacity, avour and odour com-pounds (volatiles organic compounds VOC). Spoilage commonlymanifests itself as off-odours and off-avours due to the presence Corresponding author. Tel.: +33 140966161; fax: +33 140966475.

E-mail address: fatou-toutie.ndoye@irstea.fr (F.T. Ndoye).

Journal of Food Engineering 148 (2015) 4352

Contents lists availab

od

lsfoods at all stages of the cold chain have always been among themain concerns in order to reduce food losses and health hazards.Microbial and physico-chemical quality changes may occur in foodproducts according to their timetemperature history, but also

RTE products of particular concern for food bu(FBO) and for competent authorities, knowing thposes potential human health risks (EFSA, 2013)monocytogenes, the pH and high water activityhttp://dx.doi.org/10.1016/j.jfoodeng.2014.09.0400260-8774/ 2014 Elsevier Ltd. All rights reserved.teriumn to L.k meatunder(LAB).

ity lossthreshold value of ethanol concentration was dened in relation with a threshold count numbers of LABunder the conditions studied.

2014 Elsevier Ltd. All rights reserved.

1. Introduction

Controlling and improving the quality and safety of chilled

ogenic bacteria such as Listeria monocytogenes (L. monocytogenes)is the highest along the cold chain (EFSA, 2009; Warriner andNamvar, 2009). This makes the occurrence of L. monocytogenes inReceived 21 March 2014Received in revised form 19 September 2014Accepted 24 September 2014Available online 5 October 2014

Keywords:Listeria monocytogenesLeuconostoc mesenterodesLactobacillus sakeiReady-to-eat pork productsFood safetyFood qualityCold chainBioS), BIOSYST, KU Leuven, W. de Croylaan 42, B-3001 Leuven, Belgium

a b s t r a c t

It is of crucial importance for Ready-To-Eat (RTE) foodstuffs producers to guarantee the quality and safetyof their products under the cold chain variations related to different timetemperature proles. Experi-mental designs were used to investigate and model the effects of temperature on safety and quality attri-butes of selected RTE meat products. Three types of RTE sliced pork products (cooked ham, cooked patand smoked ham) were stored at different temperatures (5, 8, 12 and 15 C) up to 6 weeks. Microbiolog-ical and physico-chemical attributes were followed. Growth parameters of Listeria monocytogenes wereinvestigated by challenge testing for the three RTE products at the four temperatures. Two lactic acid bac-teria (Lactobacillus sakei and Leuconostoc mesenterodes) were also investigated by challenge testing butonly for cooked ham and cooked pat at 8 C. Changes in quality indicators including colour, textureand water content, water activity and water dripping were evaluated over storage time for the threeRTE products. Spoilage experiments were conducted (at 2, 8, 12, 15 C for 48 days) on cooked ham andthe production of ethanol, as a representative volatile deriving from bacterial metabolism, was correlatedto bacterial outgrowth. Growth parameters of the three strains for the given food were mathematicallymodelled and validation tests were performed for L. monocytogenes in cooked ham and cooked pat.Physico-chemical attributes were not signicantly affected by timetemperature storage. The productionof ethanol on spoiled cooked ham was related to growth of lactic acid bacteria, especially Leuconostoc. AaArial, Technical Institute for Food Industry, ACTIA centre, 250 rue Laurent Fries, 67412 Illkirch, FranceSafety and quality assessment of ready-tchain

V. Stahl a, F.T. Ndoye b,, M. El Jabri c, J.F. Le Page c, BD. Thuault c

Journal of Fo

journal homepage: www.eeat pork products in the cold

ezard a, A. Lintz a, A.H. Geeraerd d, G. Alvarez b,

le at ScienceDirect

Engineering

evier .com/ locate / j foodeng

d Enof VOC. These VOC resulted from bacterial metabolism and theirproduction is correlated to bacterial cell growth (Leroy et al., 2009).

In this context, assessment of the impact of the food cold chainon microbial growth and quality attributes of RTE pork meals hasbecome important in order to improve their shelf life and theirsafety. It is also essential to have tools and methods that allow thequality and safety of foods to be accurately evaluated. Predictivemodelling is one of these tools. Several predictive models havebeen developed in order to quantify the effect of various factorson L. monocytogenes load evolution in various food products(Rosso et al., 1996; Dalgaard and Vigel Jrgensen, 1998; Zulianiet al., 2007; Couvert et al., 2010; Mejlholm et al., 2010; Huang,2013; Polese et al., 2014). These mathematical models were basedon the different phases of bacterial growth: lag; exponential; andstationary (Buchanan, 1993). They are used through software asdecision support systems for food technologists. Modelling of qual-ity deterioration during food processing and storage has also beenextensively studied (Saguy and Karel, 1980; Labuza, 1984; VanBoekel, 2008). Kinetic modelling allows the rate of deteriorationreactions in foods to be characterized as a function of temperature.Rapid assessment of RTE products spoilage can also be achieved byVOC analysis which could constitute an alternative method tomicrobiological analyses.

There are software packages for predictive microbiology avail-able (Couvert et al., 2013; Tenenhaus-Aziza and Ellouze, 2013). Inour knowledge, no such tool has been developed that combinesboth safety and quality models of foods. This is one of the objec-tives of the European FRISBEE project (Food Refrigeration Innova-tions for Safety, consumers Benet, Environmental impact andEnergy optimisation along the cold chain in Europe) which aimsto provide new tools and concepts for improving refrigerationtechnologies. The project has developed novel and innovative

Nomenclature

Latin symbolsa colour parameter; Hunter redness value ()aw water activity ()b colour parameter; yellowness value ()L colour parameter; lightness value ()Lag lag time (h)N population size (CFU)t time (h)T temperature (C)

44 V. Stahl et al. / Journal of FooQuality and Energy/Environment Assessment Tools (QEEAT) thatcombine food quality and safety together with energy and environ-mental aspects. These tools were based on mathematical model-ling as well as on a knowledge database gathering measured dataof the cold chain performance, among others, in terms of food tem-perature, and change of quality and safety attributes.

This work is part of the development of QEEAT and focuses onthe following three representative RTE pork meat products:cooked ham (a RTE cooked meat product in which the originalstructure is still recognisable), cooked pt (a RTE cooked andmixed meat product) and smoked ham (a RTE meat product inwhich the original structure is also still recognisable). The objec-tive was to develop kinetic models for safety and quality of theselected food products. The paper presents the protocols andexperiments implemented in order to develop kinetic models forbacterial growth and quality attributes changes. It presents alsoa potential alternative method to plate counting for detection ofmicrobial spoilage.2. Materials and methods

Microbial growth experiments were performed in real productsfor selected bacterial isolates (L. monocytogenes and LAB ora) as afunction of storage time and temperature. Physico-chemical attri-butes such as texture, drip-loss and colour were also experimen-tally determined. In addition, experiments were carried out torelate volatiles production to microbial population growth accord-ing to storage duration and temperature, in order to study a poten-tial, alternative method for detection of microbial spoilage ofproducts which has the advantage of being faster than classicalplate counting.

2.1. Ready-to-eat pork meals products

Three types of RTE meat products were studied: (i) cooked ham,(ii) pt, (iii) and smoked ham. They were under modied atmo-sphere packages (Pothakos et al.) with 50% CO2 + 50% N2 gasmixture.

2.2. Microbiological study

2.2.1. Bacteria strainsThe selected strain of L. monocytogenes (n 352 Arial (F)) was

isolated from environment of a meat industrial plant and was a ref-erence strain of the French program in predictive microbiology:SymPrevius (Couvert et al., 2010). Two strains were chosen forLAB, Lactobacillus (Lb.) sakei (n 1322 Aerial) and Leuconostoc mes-enterodes (N 74 Aerial), provided from the Aerial collection(stored at 80 C), isolated from chilled pork meat products.

Greek symbolsc gamma factors ()l bacterial growth rate (h1)

Subscriptsmax maximalmin minimalopt optimal0 initial

gineering 148 (2015) 43522.2.2. Optical density measurementsMeasurements of optical density allow the cardinal tempera-

ture values to be estimated. They describe the effect of tempera-ture on the growth through: (i) the minimal temperature forgrowth (Tmin); (ii) the optimal temperature for growth (Topt) and(iii) the maximal temperature for growth (Tmax) (Neysens and DeVuyst, 2005). The methodology used was adapted from Membret al. (2005) to the two LAB strains using Elliker medium (Biokar)for incubation. Turbidity growth curves of the strains were gener-ated with a Bioscreen C reader (Labosystemes Honeycomb 2France). Eleven static temperature levels between 2 C and 40 Cwere studied; pH and aw of the medium were kept at optimal lev-els, namely at 6.8 and at 0.98 respectively.

2.2.3. Microbiological challenge-testsChallenge tests in the food products were performed for the

strain of L. monocytogenes and for the two strains of LAB, using amethod adapted from Augustin et al. (2011):

one batch for each RTE product. One physiological state was

provides a good accuracy using only four descriptive parameters:lag time (lag), maximum specic growth rate (lmax), initial popula-tion size (N0) and the maximum population size (Nmax).

lnN lnN0 for t 6 laglnNmax ln 1 Nmax 1

elmaxtlag

for t > lag

(

used in order to evaluate the difference between the estimated

d Engineering 148 (2015) 4352 45used in this study: cells at the end of the exponential growthphase. Two subcultures were grown in Brain Heart Infusion(BHI) at 37 C for 16 h and for 8 h. A third subculture was car-ried out at 8 C for 6 days, in BHI broth to obtain 9 Log CFU ml1.Samples of 10 g of product were inoculated with 0.5 ml of thediluted last subculture to obtain an initial concentration of2 Log CFU g1 on the surface of the food product. Controlsamples without inoculation with L. monocytogenes were alsoincluded. Natural lactic acid ora (1 ml was plated onto deManRogosaSharpe (MRS) (Oxoid)); NF V 04-503, and aerobicmicroorganisms (1 ml was plated onto PCA (Oxoid), at 30 C for3 days), pH (Hanna HI 213; NF V04-108), and aw (GBX-FA-St 1;ISO 21807: 2004) were quantied at 3 days during challenge test-ing. Hereto, 10 g of the inoculated samples were homogenizedwith 90 ml of tryptone salt solution using a stomacher blender.Ten-fold serial dilutions were carried out and 0.1 ml of the appro-priate dilution was plated onto Compass L. monocytogenes (37 C;48 h) (Biokar): the enumerations of L. monocytogenes were per-formed on three samples at 11 different times during the lag,the exponential and the stationary phases of the growth curve.In total, twelve growth curves of L. monocytogenes in mixed-culture (with the potential presence of natural ora) wereobtained.

LAB challenge-testing were performed with one batch of cookedham and one batch of cooked pt under modied atmosphere(50% CO2 + 50% N2). Three growth curves were obtained at 8 Cin pure culture for each strain and each food product, inoculatedseparately. To allow this pure culture tests the food productswere ionized beforehand by 7.5 kGy. Three subcultures of Lb.sakei 1322 and Lc. mesenterodes 74 strains were grown at30 C; 30 C and 8 C in Elliker media. Samples of 10 g of prod-uct were inoculated with 0.5 ml of the diluted subculture toobtain an initial concentration of 2 Log CFU g1 in the foodstuff.Lb. sakei 1322 and Lc. mesenterodes 74 were quantied at regu-lar time intervals to obtain data points representing lag, expo-nential and stationary phases. The 10 g samples werehomogenized with 90 ml of tryptone salt solution (Oxoid).Ten-fold serial dilutions were carried out, and 0.1 ml of theappropriate dilution was plated onto MRS: for the enumerationof both strains. Control samples without inoculation were alsoincluded. pH and aw were measured.

2.2.4. Mathematical growth modelling of bacteria2.2.4.1. Cardinal temperature values. Cardinal temperature values ofL. monocytogenes were obtained from Couvert et al. (2010). Thecardinal values of Lc. mesenterodes 74 and Lb. sakei 1322 were esti-mated from growth kinetics obtained by experimental assays, asdescribed in Section 2.2.2.

A secondary model that described the inuence of temperatureon the growth rate (Rosso et al., 1995; Pinon et al., 2004) wasemployed to estimate the cardinal values of temperature T (Eq.(1)):

cTTTmaxTTmin2

ToptTminToptTminTTopt ToptTmaxToptTmin2T for Tmin< T < Tmax

0 Otherwise

(

1

2.2.4.2. Modelling growth kinetics. The growth kinetics in the threestudied RTE meat products were acquired at static levels of tem- L. monocytogenes challenge testing was carried out at 5, 8, 12and 15 C, under modied atmosphere (50% CO2 + 50% N2) in

V. Stahl et al. / Journal of Fooperature (5, 8, 12, 15 C). The evolution of the population size Nas a function of time was described by the logistic model withdelay (Eq. (2)) (Rosso et al., 1996; Pinon et al., 2004). This modeland predicted values.

2.3. Physico-chemical quality attributes measurements

Evolution of several quality attributes with storage time andtemperature were investigated. Colour, texture, water content,water activity and water dripping measurements were performedfor each RTE product (pt, cooked ham, smoked ham) depending

Table 1Cardinal temperature values (T, C) of Lc. mesenterodes 74 and Lb. sakei 1322estimated in liquid microbiological media.

Cardinal T (C) values Lc. mesenterodes 74 Lb. sakei 1322

T 0.53 0.25N0

2The differential form of this equation is given as following (Eq.

(3))

dNdt 0 t 6 lagdNdt lmax N 1 NNmax

t > lag

8