33.4.2 Detection of LABcontaminationin the production of vacuum-
packaged,slicedcookedham
Detection of LABcontamination sitesduring the manufacture of cookedmeat
products is difficult.LABpopulationsare verylow in an adequately cleaned
production environment(Ma»kela» and Korkeala,1987).Shortly afterpackaging,
LABpopulationsof the productsare usuallybelowthe routinedetectionlimit
(<10cfu/g),evenif the productlaterspoils quickly (Bjo»rkrothand Korkeala,
1996;Nerbrinkand Borch, 1993).The cookingprocess(internaltemperatureof
68±73ÎC) inactivates LAB and surfaces of the cooked products can be
consideredsterile(Allenand Foster, 1960;Ma»kela» and Korkeala,1987;Ma»kela»
et al., 1992). Products mainly become recontaminated with LAB during
handling afterthe cooking process.
The first studies employing DNA typing techniques for bacterial
contamination analysisassociated withfood processingwere carriedout in
orderto understandLABspoilage contaminationin the production of a vacuum-
packaged,slicedcookedhamproduct (Bjo»rkroth and Korkeala,1996,1997).
Prior to thesestudies, moleculartyping methodshad onlybeenusedin the
context of molecular epidemiological studies dealing with the spread of
pathogenicbacteriawithinpatients and the hospital environment. Molecular
typing (ribotyping) was applied in two in-plantLABcontamination analyses in a
situation where the product was expected to retain goodsensorial qualityfor 21
days,but souroff-odours and off-flavours were sometimesdetected after 14
days.The aim of thesestudieswas to locatethe sitesand sources of spoilage
LABcontamination.
In the firstphaseof the study, a Lactobacillussakeistarter strainwas
suspectedof contaminating and spoiling the product duringthe slicingand
vacuumpackagingsteps.Thisspecieswas knownto causespoilage in cooked
meat products(Korkeala and Bjo»rkroth, 1996)and cross-contaminationbetween
a fermentedmeat product and the cookedhamwas suspected.A totalof 127
LABstrains fromproducts, packagingroomsurfaces and the fermentbatches
weretyped usingEcoRI andHindIII rRNAgeneRFLPanalysis (Bjo»rkroth and
Korkeala,1996).These patternswereusedto recognizethe starter strainamong
the contaminatingLABdetectedin the productand manufacturingenvironment.
This analysisshowed that the fingerprint of the L. sakei starter was not
associated with the spoilage of ham.Therefore, the second contamination
analysis (Bjo»rkroth and Korkeala,1997)was carried out.
A total of 982LAB isolates from the raw materials, product and the
environmentat differentproduction stageswerescreened. LABtypingresulted
in 71 differentribotypes,of which27 wereassociated withcontamination
routes. Raw material (macerated porkmeat)was distinguishedas the sourceof
the majorspoilage strains. Contamination of the productsurfaces aftercooking
was shown to be airborne. The removalof the product fromthe cooking-forms
was localizedas a majorsite for airborneLABcontamination.Foodhandlers
and somesurfacesin contactwiththe product duringthe manufacture werealso
contaminatedwiththe spoilage strains.SomeLABstrainswereshown to be able
548 Handbookof hygiene controlin the foodindustry