1260

Journal of Food Protection Vol 59 No 12 1996 Pages 1260-1266Copyright copy International Association of Milk Food and Environrnantal Sanitarians

Validation of Pepperoni Processes for Controlof Escherichia coli 0157H7+

JAY C HINKENSlt NANCY G FAITH2 TIMOTHY D LORANG PHILLIP BAILEYlDENNIS BUEGE3 CHARLES W KASPAR2 and JOHN B LUCHANSKy24

IDoskocil Companies 1nc Jefferson Wisconsin 53549 and 2Department of Food Microbiology and Toxicology Food Research 1nstitute3Muscle Biology Laboratory and4Department of Food Science University of Wisconsin Madison Wisconsin 53706 USA

(MS 96-18 Received 7 February 1996Acceptedl April 1996)

ABSTRACT

The outbreak of Escherichia coli 0157H7 linked withdry-cured salami in late 1994 prompted regulatory action thatrequired manufacturers of fermented products to demonstrate a5-log unit reduction in counts of this pathogen during processingTherefore pepperoni batter (75 pork25 beef with a fat contentof ca 32) was inoculated with a pediococcal starter culture and afive-strain mixture of E coli 0157H7 (~2 X 107 CFUg) andstuffed into 55-mm diameter fibrous casings 47 cm in length Theviability of the pathogen was monitored before stuffing afterfermentation after thermal processing andor after drying Chubswere fermented at 96degF (36degC) and 85 relative humidity (RH) topH 0 50 and then dried at 55degF (l3degC) and 65 RH to amoistureprotein ratio of 0161 (modified method 6 process)Counts of the pathogen decreased about 12 log units afterfermentation and drying In subsequent experiments heating chubsafter fermentation to internal temperatures of 145degF (63degC) instan-taneous or 128degF (53degC) for 60 min resulted in a ~5-log unitdecrease in numbers of strain 0157H7 without visibly affectingthe texture or appearance of the product These data revealed that atraditional nonthermal process for pepperoni was only sufficienttoeliminate relatively low levels (ca 2 log CFUg) of E coli0157H7 whereas heating to internal temperatures of 145degF(63degC) instantaneous or 128degF (53degC) for 60 min delivered a 5 to 6log unit reduction in counts of the pathogen in pepperoni

Key words Escherichia coli 0157H7 pepperoni pathogenfermentation

The Western United States outbreak of Escherichia coli0157H7 in 1993 due to consumption of contaminatedhamburger patties at a fast-food restaurant chain (7) broughtconsiderable focus on this pathogen Despite the heightenedawareness approximately 1 year after the hamburger out-

Author for correspondence Tel 608-263-7777 Fax 608-263-1114Email jbluchanfacstaffwiscedu

t Present Address Silliker Laboratories of Wisconsin 3688 Kinsman BlvdMadison WI 53704

t Portions of this research were presented at the Annual Meeting of theInternational Association of Milk Food and Environmental Sanitarians30 June to 3 July 1996 in Seattle Washington USA( 14)

break another outbreak of E coli 0157H7 was linked tomeat this time a dry fermented pork and beef salami (8)The salami outbreak involved 20 individuals in Washington(age range 23 months to 77 years median age 6 years 3hospitalized one 6-year-old with hemolytic uremic syn-drome [HUS]) and 3 individuals in California (age range 4to 75 years 3 hospitalized one 4-year-old with HUS) (8) Ecoli 0157H7 was found on presliced product from delicates-sen counters (8) In early 1995 there was another outbreak ofHUS linked to semidry (uncooked) fermented sausage (9)This outbreak involved 23 patients (age range 4 months to12 years median age 4 years 23 with HUS) in SouthAustralia and was attributed to E coli 0111 NM anotherserotype of E coli that produces a Shiga-like toxin(s) As aresult of the California and Washington salami outbreak theUS Department of Agriculture (USDA) Food Safety andInspection Service (FSIS) developed guidelines (24) forsausage manufacturers to validate processes to ensure a5-log unit reduction in counts of E coli 0157H7

Among the many different categories of fermentedsausage pepperoni is one of the most popular varieties withannual consumption in the United States exceeding 370million pounds (Gerry Durnell personal communiciltion)Pepperoni types and sizes vary greatly but three basiccategories predominate First there is the traditional small-diameter (28 to 36 mm) product found in the deli case or onpegboards in supermarkets The second type representingthe largest volume consumed is pepperoni stuffed into 49-to 54-mm diameter casings and then sliced (1 to 2 mm thick)and sold to pizza restaurants A small portion of this type ofproduct is also sold as sticks and subsequently used as atopping for frozen pizza The third category of pepperoni isthe large-diameter (60 to 80 mm) product used on sand-wiches Differences among manufactures and categories ofpepperoni notwithstanding the use of various combinationsof salts pH heating cooling andor drying ensure thatpepperoni is wholesome and safe relative to the parasiteTrichinella spiralis and to bacterial pathogens such asStaphylococcus aureus In contrast to date there has beenno information available to certify that pepperoni is safe

E COLI 0157H7 IN PEPPERONI 1261

relative to E coli 0157H7 To this end the objective of thepresent study was to validate the effect of different process-ing conditions on the fate of E coli 0157H7 in pepperoni

MATERIALS AND METHODS

Bacterial strainsFive strains of E coli 0157H7 were used in this study to

inoculate pepperoni (i) strain SLH21788 human isolate from theWisconsin daycare outbreak of 1994 (22) (ii) strain C7927 humanisolate from the Massachusetts apple cider outbreak of 1991 (6)(iii) strain F-90 sausage isolate from the California dry-curedsalami outbreak of 1994 (8) (iv) strain EC204P pork isolate fromthe Food Research Institute Culture Collection and (v) strainC9490 human isolate from the Western States hamburger outbreakof 1993 (7) The E coli strains were maintained at -70degC in 2XTrypticase soy broth (TSB) (Difco Laboratories Inc Detroit MI)plus 10 glycerol Prior to inoculation into pepperoni batter thecultures were transferred twice in TSB supplemented with 1glucose at 37degC overnight before use as outlined by the USDAFSIS (24) A commercial Pediococcus acidilactici starter culture(Lactacel 115 Quest International Sarasota FL) was maintainedand propagated according to the manufacturers instructions

Preparation ofE coli 0157H7 inoculumEach of the five strains of E coli 0157H7 were separately

grown overnight in 250 ml of TSB plus I glucose at 37degC withshaking (100 rpm) The cells were harvested by centrifugation at1880 X g for 20 min at 4degC and washed and resuspended in asmall volume of 01 peptone (Difco) The five cell suspensionswere combined and the final volume was adjusted to 50 ml with01 peptone such that each strain contributed about 20 to thefinal inoculum of 1 X 109 CFUml The five-strain cocktail washeld on ice for ~30 min before addition to sausage batter asdescribed below The viable count of the cocktail was confirmed byspread plates on MacConkey sorbitol agar (MSA) (Difco) just priorto inoculation of the batter

Manufacture of pepperoniA flow diagram for sausage manufacture is provided in Figure

1 The meat block supplied by Doskocil Companies (Jefferson WI)was maintained at OdegC during batter preparation The batter (ca 25Ib [lIb is ca 045 kg] per trial 75 [wtwt] pork and 25 [wtwt]beef) with a target fat content of ca 32 (wtwt) contained 063

BatterJ

InoculationJ

Fermentation(96degF 85 RH pH 050)

~Dry Heat

(55degF65RHMPr161) ~

t 145F 128F60

l__ I__ 1

FIGURE 1 Pepperoni manufacture Flow diagram depicting thepepperoni processes evaluated MIPr moisture to protein ratio

dextrose (016 Ib25 Ib of batter A E Staley Decatur IL) 3spice mix (075 Ib251b of batter Doskocil house blend) and 20cure mixture (05 Ib25 Ib of batter Doskocil house blend) Thecure and spice mixtures were formulated to contain garlic anisered pepper clove and oleoresin of paprika BHA (28 ppmEastman Chemical Kingsport TN) BHT (butylated hydroxytolu-ene) (28 ppm Eastman) and citric acid (28 ppm Fuyang Pharma-ceutical China) and to provide a final batter concentration of 156ppm NaN02 and ca 333 NaCl The batter was inoculated withthe five-strain cocktail of E coli 0157H7 (final concentration of22 X 107 CFUg of batter) and mixed using a Buffalo mixer(model 2VSS John E Smiths and Sons Co Buffalo NY) for 2min Six milliliters of thawed pediococcal starter culture in 42 ml ofsterile dH20 were added per 25 Ib of batter to deliver ca 108 CFUgof batter The commercial spice and cure were added and mixingwas continued for an additional 2 min A noninhibitory food gradegreen dye (FDampC yellow 5 and FDampC blue 1 McCormick amp CoInc Hunt Valley MD) was added to the batter as an aid to monitorfor uniform mixing A control batter was prepared prior to theexperimental batter that was identical to the latter except for thepresence of the serotype 0157H7 cocktail Chemical analyseswere performed (see below) on representative chubs prepared fromthe control batter The inoculated batter was removed from themixer and ground through a 0125 in (ca 032 cm) die plate using aHobart grinder (model 84142 Hobart Manufacturing Co TroyOH) The batter was then stuffed (ca 694 g per chub) using a handstuffer (Koch Supplies Inc Kansas City MO) into 55-mmdiameter fibrous casings (TeePak Inc Westchester IL) to a finallength of 47 cm and the resulting chubs were sealed using a casingclipper (Poly-clip model SCH 7210 Niedecker West Germany)loaded with series 7000- VSCC staples (US Clip CorporationMundelein IL) Casings were soaked in tap H20 at room tempera-ture for 10 min prior to use

The chubs were transferred to a smokehouse (model 1000Vortron Inc Beloit WI) and fermentation was conducted at 96degF(36degC) with a relative humidity (RH) of 85 to 90 until reaching apH ~ 50 After fermentation (ca 14 to 18 h) the chubs werecold-showered to an internal temperature of ~80degF (27degC) in thesmokehouse and then dried at 55degF (13degC) and 65 RH [wet bulb50degF (10degC) dry bulb 55degF (13degC)] to a moistureprotein (MPr)ratio of 16 1 (ca 15 to 21 days) in an environmentally controlledchamber (Biotron facility University of Wisconsin Madison WI)In addition to the modified method 6 (27) process just describedfor some experiments chubs were heated to internal temperatures of14SOF (63degC) instantaneous or 128degF (53degC) for 60 minutes andcold-showered before drying Three trials were conducted for eachof the three processes evaluated (i) modified method 6 (ii)modified method 6 with a postfermentation heating of chubs to aninternal temperature of 145degF (63degC) and (iii) modified method 6with a postfermentation heating of chubs to an internal temperatureof 128degF (53degC) for 60 min

Microbiological analyses of pepperoni chubsPepperoni (two chubs per sampling point) was tested for

viable E coli 0157H7 by direct plating prior to stuffing afterfermentation after cooking (when applicable) and after 4 8 1218 and 21 days of drying When numbers of the pathogendecreased below detection by direct plating laquo 102 CFUg) thepresence or absence of the pathogen was determined by enrich-ment The noninoculated batter was also tested for backgroundlevels of E coli 0157H7 and other non-sorbitol-fermentingbacteria by spread plating onto MSA and for total aerobic bacterialnumbers by spread plating onto Trypticase soy agar (TSA) (Difco)plates Immediately after stuffing the chubs were also tested for

1262 HINKENS FAITH LORANG BAILEY BUEGE KASPAR AND LUCHANSKY

viable pediococci by spread plating onto MRS (Difco) agar platesFor sampling chubs the outside of the chub was cleaned with 70ethanol and the casing was removed with the aid of a sterile scalpelA 25-g portion of the batter or a 25-g cross-section from the middleof a chub was asceptically transferred to a Stomacher bag (SewardMedical London UK) containing 225 ml of 01 peptone and thecontents stomached (model 400 Telanar Co Cincinnati OH) for 2min at room temperature One milliliter of the resulting mixturewas serially diluted in 01 peptone and spread plated ontoduplicate MSA or MRS agar plates Another portion (10 ml) wastransferred to 90 ml of modified EC broth (21) containingnovobiocin (final concentration of 20 figml Sigma Chemical CoSt Louis MO) and incubated with shaking (100 rpm) for 18 to 24 hat 37degC to test for the presence or absence of low levels of viable0157H7 cells Following enrichment the samples were diluted in01 peptone and spread plated in duplicate onto MSA Plateswere incubated at 42degC for 48 h to recover E coli 0157H7 and at37degC for 24 h to recover pediococci before colonies were counted

The serotype of six sorbitol nonfermenting (ie white)colonies from MSA plates was determined using an 0157 latexagglutination test (Unipath Limited Basingstoke England) Sorbi-tol-negative 0157-positive isolates were further confirmed as Ecoli using API 20E biochemical test strips (bioMerieux Vitek IncHazelwood MO) and for the H7 antigen by agglutination (Difco)

Chemical analyses of pepperoni chubsAt each sampling point two control chubs (without E coli

0157H7) were removed for sampling The control chubs wereeither transported on ice directly to the Doskocil Companies(Jefferson WI) for testing or held at -20degC for up to seven daysand then delivered to Doskocil Chemical analyses were performedfollowing Association of Official Analytical Chemists (AOAC)procedures (19) Each chub was tested for fat (AOAC procedure96039) moisture (AOAC procedure 95046) protein (AOACprocedure 92806) and salt (AOAC procedure 93547) The wateractivity (aw) was determined using a Rotronic water activity meter(model DT Huntington NY) The pH and titratable acidity (TA)were determined as previously described (18) using a 25-g sampleof sausage macerated in 100 ml of hot (ca 60degC) dH20 The TAwas expressed as the percent lactic acid

Statistical analysisData were analyzed using the Statistical Analysis System

(SAS Institute Cary NC)

RESULTS

Microbiological testing of raw meat and batterAnalyses of meat blocks before inoculation with the

five-strain cocktail and pediococcal starter culture revealedthat none of7 meat blocks tested contained E coli 0157H7by direct plating (data not shown) In addition to counts ofserotype 0157H7 a portion of these same 7 meat blockswas also tested to determine the total aerobic plate countLevels ranged from lt 102 to 4 X 104 CFUg of meat(average 13 X 104 CFUg) indicating that the raw materi-als were of excellent microbiological quality

Prior to addition of the pediococcal starter culture theantimicrobial capabilities of the spice and cure mixtures onthe five-strain mixture of E coli 0157 H7 was also ascer-tained The data revealed that the spice and cure mixturesused in this study had no significant effect on the numbers of0157H7 (a lt 005 P = 09945) The average count of Ecoli 0157H7 from emulsions prepared from 6 of the 7 meatblocks tested prior to the addition of spice and cure was25 X 107 CFUg of batter (range 11 X 106 to 14 X 108CFUg) The average count from these 6 emulsions testedafter the addition of the spice and cure mixtures and grindingwas 22 X 107 cfu per gram of batter (range 15 X 105 to13 X 108 cfugram)

Fermentation at 96degF (36degC) to pH 50 and drying at55degF (BdegC) and 65 RH to a moistureprotein (MPr)ratio of161

The first manufacture process evaluated was the modi-fied method 6 process for pepperoni (Table 1) wherein thepathogen cell numbers remained essentially the same duringthe 14- to 18-h fermentation Although a 12 log CFUg

TABLE 1 E coli 0157H7 count chemical composition and heating regimen of pepperoni during fermentation at 96degF to pH lt50 anddrying at 55degF and 65 RH to a moisture to protein ratio of 5161

Chemical analysis (value SD 11 = 3) for

Sampling time

E coli0157H7(log CFUg SD) pH TA()a Salt () Moisture () Protein () Fat ()

659 29 472 09 77 03 259 Il 95 00 363 09 4177 060 1613 61 3810 220650 47 464 08 86 04 21212 93 01 383 17 3717 116 1757 49 3973 09261437 473 05 88 04 178 17 91 00 377 12 3273 217 1843 97 4293 100597 38 472 09 88 04 153 17 89 00 390 00 2943 189 1933 82 4500 108569 23 474 09 93 09 135 14 87 02 440 29 2747 197 2033 73 4660 144

Day 0Emulsion 689 86Afterfermen- 682 24 48517 62 08 346 07 96 00 317 09 4883 046 1407 plusmn 29 3113 229tatione

Drying day48121821

a TA titratable acidity as lactic acidb MPr moisture to protein ratioC Smokehouse settings were dry bulb 85degF and wet bulb 80degF during conditioning for I h and dry bulb 96degF and wet bulb 92degF duringfermentation to SpH 50

E COLI 0157H7 IN PEPPERONI 1263

a TA titratable acidity as lactic acidb MPr moisture to protein ratioC No positiveno of chubs tested by enrichmentd Smokehouse settings were dry bulb 85degF and wet bulb 80degF during conditioning for 1 h and dry bulb 96degF and wet bulb 92degF duringfermentation to pH 550

e Postfermentation settings were dry bulb 115degF and wet bulb 104degF for 1 h for cycle 1 dry bulb 120degF and wet bulb 102degF for I h for cycle2 dry bulb 125degF and wet bulb 115degF forI h for cycle 3 dry bulb 135degF and wet bulb 126degF for 1 h for cycle 4 and dry bulb 157degF and wetbulb 148degF until 145degF internal temperature reached for cycle 5

decrease was observed after 21 days of drying the desired5-1og CFUg reduction was not achieved The proximatecomposition of the pepperoni displayed the expected levelsfor pH aw salt protein moisture and fat (Table 1) Inrelated experiments the dextrose concentration was adjustedfrom 063 to 090 to extend the fermentation to 24 handlower the pH to lt47 As in the 14- to 18-h fermentationcounts of the pathogen only decreased ca 05 log CFUgafter fermentation and another 141 log CFUg after 12 daysof drying in the single trial conducted (data not shown)These data revealed that the modified method 6 process forpepperoni was not sufficient to eliminate high levels of

serotype 0157H7 strains of E coli following 14 to 24 h offermentation and subsequent drying

Fermentation at 96degF (36degC) to pH5 50 heating to aninternal temperature of 145degF (63degC) and drying at 55degF(13degC) and 65 RH to a MProf5161

Since the modified method 6 process did not produce a5-D kill as required by the USDAFSIS experiments wereconducted to evaluate postfermentation heating for eliminat-ing the pathogen In the first of two such experiments afterreaching pH 5 50 pathogen numbers remained essentiallythe same as starting levels (Table 2) The smokehouse air

a TA titratable acidity as lactic acidb MPr moisture to protein ratioC No positiveno of chubs tested by enrichmentd Smokehouse settings were dry bulb 85degF and wet bulb 80degF during conditioning for 1 h and dry bulb 96degF and wet bulb 92degF duringfermentation to pH 550

e Postfermentation settings were dry bulb 106degF and wet bulb 98degF for I h for cycle 1 dry bulb 120degF and wet bulb 102degF for 1 h for cycle 2dry bulb 130degF and wet bulb 115degF for 1 h for cycle 3 dry bulb 140degF and wet bulb 123degF for I h for cycle 4 and dry bulb 140degF and wetbulb 130degF until 128degF internal temperature reached and then hold for 60 min for cycle 5

1264 HINKENS FAITH LORANG BAILEY BUEGE KASPAR AND LUCHANSKY

temperature was then raised at a rate of 12 to 15degF (70 toSoC) per h until an internal chub temperature of 145degF(63degC) was reached (Table 2) Heating to 145degF (63degC)reduced counts of the pathogen below detection (gt50-logunit decrease) however it was possible to recover viablecells of serotype 0157 H7 by enrichment after heating andup to S days of drying Nevertheless heating chubs to aninternal temperature of 145degF (63degC) delivered the required5-log unit decrease in counts of the pathogen withoutcausing any visible textural or color changes to the finalproduct (data not shown)

Fermentation at 96degF (36degC) to pHS 50 heating to aninternal temperature of 128degF (53degC) for 60 min anddrying at 55degF (13degC) and 65 RH to a MIProfS161

On the basis of the encouraging results with the 145degF(63degC) heating process a second heating regimen wasevaluated As in the previous experiments fermentationalone had no appreciable effect on the pathogen (09-logcfug decrease) (Table 3) However holding the chubs at aninternal temperature of 12soF (53degC) for 60 min (Table 3)resulted in at least a 5-log CFUg reduction in numbers of0157H7 After heating it was only possible to recover thepathogen by enrichment Also heating for 60 min at 12soF(53degC) did not visibly affect the appearance of the resultingproduct and did not cause greasing out (ie there was noresidual fat layer under the casing) (data not shown)

DISCUSSION

Escherichia coli 0157H7 emerged as a foodbornepathogen of primary health concern in the 1990s During thisperiod serotype 0157H7 strains were incriminated innumerous sporadic cases and several outbreaks involving avariety of foods (6 7-9 22 29) Foods of bovine origin ingeneral and ground beef in particular were most oftenidentified as the food vehicles (16) To better understand theresponse of this bacterium to intrinsic (eg pH) andextrinsic (eg temperature) conditions associated with foodsstudies were conducted using apple cider (6 20) cottagecheese (4) ground beef (13 17 26) mayonnaise andordressings (23 29) and salad vegetables (1) In general thesestudies revealed E coli 0157 H7 is very tolerant to acidAlthough there is significant variation in the degree of acidtolerance from strain to strain most 0157H7 strains aremore acid tolerant than other E coli (3 5 11 20) HoweverE coli 0157H7 does not appear to be particularly heattolerant (13) Although there have been reports on thebehavior of non-0157H7 strains of E coli in fermentedmeats (12 28) with the exception of a previous studyconducted at the Food Research Institute using salami (15)no information has been published on the behavior ofserotype 0157H7 strains in fermented sausages

Several factors associated with pepperoni batter wereevaluated for possible effects on the viability of the serotype0157H7 cocktail For example high levels of indigenousbacteria could have a negative impact on the behavior of Ecoli 0157H7 However the relatively low numbers ofaerobic bacteria in the raw materials suggest that indigenous

bacteria were probably not a significant factor in this studyLikewise microbiological analyses of the batter revealedthat the spice and cure mixtures had no significant effect onthe 0157H7 cocktail Due to the limitations of the facilityhousing the smokehouse it was not possible to evaluate theeffect of smoke in this study Smoking is presumably mosteffective against surface contamination of sausage and assuch was not expected to have an appreciable effect on cellsof E coli within pepperoni chubs The addition of smokecould enhance the kill achieved by fermentation but itwould probably not be sufficient to deliver a 5-D killAmong the variety of other factors that could influence theviability of serotype 0157H7 strains in pepperoni the pHand acid type as well as the time held at a given temperatureprobably present the greatest potential for impacting patho-gen numbers (2) For example E coli 0157H7 is moresensitive to acid conditions (pH 36 to 40) at abuse (15 to20degC or 60 to 77degF) rather than refrigeration (4 to SoC or 39to 46degF) temperatures (6 20) It is also significant thatsurvival of serotype 0157H7 strains can also be affected bythe type of acid In a study by Abdul-Raouf et al (2)evaluating the effect of acidulants on E coli 0157H7 inbeef slurries the order of effectiveness for inhibiting growthwas acetic gt lactic gt citric acid at all pH values andtemperatures tested However Conner and Kotrola (10)reported that organic acids and salts enhanced the survival ofserotype 0157H7 strains during storage at 4degC in asynthetic medium These data underscore the importance ofconducting validation studies in bona fide fermented meatsrather than microbiological media andor beaker sausage tomore precisely establish the viability of E coli 0157H7 fora given formulation and manufacturing process

Although lower pH andor higher fermentation tempera-tures may result in a greater decrease in pathogen numbersit is unlikely that these parameters could be adjustedsufficiently to deliver a 5-D kill in pepperoni More specifi-cally at pH S 43 pepperoni may be too tart for mostconsumers Similarly due to intrinsic limitations of thestarter culture it may not be possible to conduct a fermenta-tion at temperatures much in excess of 120degF (49degC) Forthese reasons it was necessary to evaluate postfermentationheating of chubs as a method to deliver the mandated 5-Dreduction in counts of E coli 0157H7 The first thermalprocess tested 145degF (63degC) internal was selected becauseit meets the requirements for trichinae destruction andbecause it approximates the time temperature and humidityguidelines established by the USDA for cooked andor roastbeef (27) The cooking requirements for cooked andor roastbeef were also approved by the USDA FSIS as an option tocontrol serotype 0157H7 strains of E coli in fermentedsausage (25) The other thermal process tested 12soF(53degC) internal for 60 min was selected because it alsomeets the requirements for trichinae control but moreimportant it is less harmful to the texture and appearance ofpepperoni than the 145degF (63degC) cooking regimen (SteveSeideman personal communication) This approach is alsonot without limitations as some sausage varieties are notamenable to postfermentation heating due to possible unde-

E COLI 0157H7 IN PEPPERONI 1265

sirable textural effects on the product andor some manufac-turers are not capable of including such a heating step due tothe volume of product produced or the cost of the attendantequipment and facilities From a practical standpoint heat-ing may cause greasing out and may yield a product moreprone to cupping when used as a pizza topping (SteveSeideman personal communication) Further studies may bewarranted to identify heating regimens that are sufficient toeliminate the pathogen without producing untoward effectson the finished product and that are cost effective

The present study evaluated the modified method 6pepperoni process as well as postfermentation heating toinactivate E coli 0157H7 A comprehensive sampling andscreening of raw materials to account for low levels of thispathogen in raw materials as well as adherence to goodmanufacturing procedures would also contribute to thesafety of fermented meat products As another considerationthe sampling plan for raw materials that will undergosubsequent heat processing such as pepperoni used as pizzatopping is less stringent (ie 15 samples per lot comparedto 30 samples per lot [25 J) because the ensuing cooking ispresumably sufficient to eliminate the pathogen Perhaps amore serious threat to consumer safety would result fromcontaminated pepperoni used for sandwiches or salad barsFor these uses it is necessary to include a postfermentationheating step sufficient to deliver the 5-D kill Thus it wouldalso be informative to substantiate the fate of the pathogenafter frozen or refrigerated storage of pepperoni targeted foruse in sandwiches on pizza or for salad bars

The results of the present study revealed that thenonthermal process (modified method 6) for pepperonimanufacture was not sufficient to effect a 5-10g unit reduc-tion in counts of E coli 0157H7 The observed lt2-10g unitdecrease in counts of the pathogen following fermentation ofpepperoni at 96degF (36degC) to pH ~ 50 and drying was inagreement with a previous report of a 1- to 2-10g unitdecrease in counts of the pathogen under similar processconditions for salami (ie fermentation at 96degF (36degC) topH 48 drying to a MPr of ~191 and storage at 4degC for 2weeks) (15) In the present study postfermentation heatingachieved a 5-D kill of this pathogen Although both heatingregimens evaluated were effective for eliminating E coliOI57H7 cooking at l28degF (53degC) for 60 min was lessharmful to the product than cooking to an internal tempera-ture of 145degF (63degC) These data represent the first of severalreports expected to be published on the fate of E coli0157H7 in fermented meats by our group and by otherinvestigators In addition to a thermal processing step weare evaluating the eff~ct of fermentation temperature andpH a 3- to 7-day postfermentation storage at low pH andvarious heating regimens for eliminating this pathogen inready-to-eat fermented sausage It is possible that slightincreases in the fermentation temperature andor extendedstorage at low pH would result in a greater reduction incounts of the pathogen and further reduce the time andtemperature required for postfermentation heating A compre-hensive effort to screen raw materials for the pathogen tofollow good manufacturing practices and to monitor fin-ished product would significantly reduce the likelihood of

illness due to E coli 0157H7 in fermented sausage In theinterim the results of the present study will assist manufac-turers in prioritizing options for validating the safety of theirrespective manufacturing processes

ACKNOWLEDGMENTS

The technical assistance of Patrick Krakar Tamara Dodge and KarenArnold of the Food Research Institute and Rong Chen and Jane Wagner ofDoskocil Foods are gratefully acknowledged Our appreciation is alsoextended to Chuck Baum (OW-Madison Biotron) John Cerveny (OscarMayer) Steven Devcich (TeePak) Gerry Durnell (National Association ofPizza Operators) Larry Hand (Diversitech) Steve Lary (Alkar) and SteveSeideman (Doskocil) for supplies service equipment andor consultationsWe also thank the following individuals for supplying cultures Jay HLewis (Washington State Public Health Lab) for strain F-90 Mary EProctor (Wisconsin Bureau of Public Health) for strain SLH21788 KathyGlass (Food Research Institute) for strain C7927 and Joy G Wells andTimothy J Barrett (Centers for Disease Control and Prevention) for strainC9490

This work was supported by a gift from Doskocil Companies Inc andthe College of Agricultural and Life Sciences of the University ofWisconsin-Madison

REFERENCES

1 Abdul-Raouf U M L R Beuchat and M S Ammar 1993 Survivaland growth of Escherichia coli 0157H7 on salad vegetables ApplEnviron Microbiol 59 1999-2006

2 Abdul-Raouf U M L R Beuchat and M S Ammar 1993 Survivaland growth of Escherichia coli 0157H7 in ground roasted beef asaffected by pH acidulants and temperature Appl Environ Micro-bioI 592364-2368

3 Arnold K W and C W Kaspar 1995 Starvation- and stationary-phase-induced acid tolerance in Escherichia coli 0157H7 ApplEnviron Microbiol 612037-2039

4 Arocha M M M McVey S D Loder J H Rupnow and L BBullerman 1992 Behavior of hemorrhagic Escherichia coli 0157H7during the manufacture of cottage cheese J Food Prot 55379-381

5 Benjamin M M andA R Datta 1995 Acid tolerance of enterohem-orrhagic Escherichia coli Appl Environ Microbiol 61 1669-1672

6 Besser R E S M Lett J T Weber M P Doyle T J Barrett J GWells and P M Griffin 1993 An outbreak of diarrhea and hemolyticuremic syndrome from Escherichia coli 0157H7 in fresh-pressedapple cider JAMA 2692217-2220

7 Centers for Disease Control and Prevention 1993 Update multi stateoutbreak of Escherichia coli 0157H7 infections from hamburgers-western United States 1992-1993 Morbid Mortal Weekly Rep42258-263

8 Centers for Disease Control and Prevention 1995 Escherichia coli0157H7 outbreak linked to commercially distributed dry-curedsalami-Washington and California 1994 Morbid Mortal WeeklyRep 44157-160

9 Centers for Disease Control and Prevention 1995 Communityoutbreak of hemolytic uremic syndrome attributable to Escherichiacoli 0111 NM-South Australia 1995 Morbid Mortal Weekly Rep44550-551 557

10 Conner D E and J S~ Kotrola 1995 Growth and survival ofEscherichia coli 0157H7 under acidic conditions Appl EnvironMicrobiol 61382-385

l1 Cutter C N and G R Siragusa 1994 Efficacy of organic acidsagainst Escherichia coli 0157H7 attached to beef carcass tissue usinga pilot scale model carcass washer J Food Prot 5797-103

12 Darmadjii P M Izumimoto T Miyamoto and K Katoaka 1990Lactic fermentation on preservative qualities of Dendeng giling JFood Sci 55(6) 1523-1527

13 Doyle M P and J L Schoeni 1984 Survival and growthcharacteristics of Escherichia coli associated with hemorrhagiccolitis Appl Environ Microbiol 48855-856

14 Faith N G J Hinkens T D Lorang P Bailey D Buege C WKaspar and J B Luchansky 1996 Validation of pepperoni processesfor control of Escherichia coli 0157H7 abstr 191 p 80 Progr

E COLI 0157H7 IN PEPPERONI 1261

relative to E coli 0157H7 To this end the objective of thepresent study was to validate the effect of different process-ing conditions on the fate of E coli 0157H7 in pepperoni

MATERIALS AND METHODS

Bacterial strainsFive strains of E coli 0157H7 were used in this study to

inoculate pepperoni (i) strain SLH21788 human isolate from theWisconsin daycare outbreak of 1994 (22) (ii) strain C7927 humanisolate from the Massachusetts apple cider outbreak of 1991 (6)(iii) strain F-90 sausage isolate from the California dry-curedsalami outbreak of 1994 (8) (iv) strain EC204P pork isolate fromthe Food Research Institute Culture Collection and (v) strainC9490 human isolate from the Western States hamburger outbreakof 1993 (7) The E coli strains were maintained at -70degC in 2XTrypticase soy broth (TSB) (Difco Laboratories Inc Detroit MI)plus 10 glycerol Prior to inoculation into pepperoni batter thecultures were transferred twice in TSB supplemented with 1glucose at 37degC overnight before use as outlined by the USDAFSIS (24) A commercial Pediococcus acidilactici starter culture(Lactacel 115 Quest International Sarasota FL) was maintainedand propagated according to the manufacturers instructions

Preparation ofE coli 0157H7 inoculumEach of the five strains of E coli 0157H7 were separately

grown overnight in 250 ml of TSB plus I glucose at 37degC withshaking (100 rpm) The cells were harvested by centrifugation at1880 X g for 20 min at 4degC and washed and resuspended in asmall volume of 01 peptone (Difco) The five cell suspensionswere combined and the final volume was adjusted to 50 ml with01 peptone such that each strain contributed about 20 to thefinal inoculum of 1 X 109 CFUml The five-strain cocktail washeld on ice for ~30 min before addition to sausage batter asdescribed below The viable count of the cocktail was confirmed byspread plates on MacConkey sorbitol agar (MSA) (Difco) just priorto inoculation of the batter

Manufacture of pepperoniA flow diagram for sausage manufacture is provided in Figure

1 The meat block supplied by Doskocil Companies (Jefferson WI)was maintained at OdegC during batter preparation The batter (ca 25Ib [lIb is ca 045 kg] per trial 75 [wtwt] pork and 25 [wtwt]beef) with a target fat content of ca 32 (wtwt) contained 063

BatterJ

InoculationJ

Fermentation(96degF 85 RH pH 050)

~Dry Heat

(55degF65RHMPr161) ~

t 145F 128F60

l__ I__ 1

FIGURE 1 Pepperoni manufacture Flow diagram depicting thepepperoni processes evaluated MIPr moisture to protein ratio

dextrose (016 Ib25 Ib of batter A E Staley Decatur IL) 3spice mix (075 Ib251b of batter Doskocil house blend) and 20cure mixture (05 Ib25 Ib of batter Doskocil house blend) Thecure and spice mixtures were formulated to contain garlic anisered pepper clove and oleoresin of paprika BHA (28 ppmEastman Chemical Kingsport TN) BHT (butylated hydroxytolu-ene) (28 ppm Eastman) and citric acid (28 ppm Fuyang Pharma-ceutical China) and to provide a final batter concentration of 156ppm NaN02 and ca 333 NaCl The batter was inoculated withthe five-strain cocktail of E coli 0157H7 (final concentration of22 X 107 CFUg of batter) and mixed using a Buffalo mixer(model 2VSS John E Smiths and Sons Co Buffalo NY) for 2min Six milliliters of thawed pediococcal starter culture in 42 ml ofsterile dH20 were added per 25 Ib of batter to deliver ca 108 CFUgof batter The commercial spice and cure were added and mixingwas continued for an additional 2 min A noninhibitory food gradegreen dye (FDampC yellow 5 and FDampC blue 1 McCormick amp CoInc Hunt Valley MD) was added to the batter as an aid to monitorfor uniform mixing A control batter was prepared prior to theexperimental batter that was identical to the latter except for thepresence of the serotype 0157H7 cocktail Chemical analyseswere performed (see below) on representative chubs prepared fromthe control batter The inoculated batter was removed from themixer and ground through a 0125 in (ca 032 cm) die plate using aHobart grinder (model 84142 Hobart Manufacturing Co TroyOH) The batter was then stuffed (ca 694 g per chub) using a handstuffer (Koch Supplies Inc Kansas City MO) into 55-mmdiameter fibrous casings (TeePak Inc Westchester IL) to a finallength of 47 cm and the resulting chubs were sealed using a casingclipper (Poly-clip model SCH 7210 Niedecker West Germany)loaded with series 7000- VSCC staples (US Clip CorporationMundelein IL) Casings were soaked in tap H20 at room tempera-ture for 10 min prior to use

The chubs were transferred to a smokehouse (model 1000Vortron Inc Beloit WI) and fermentation was conducted at 96degF(36degC) with a relative humidity (RH) of 85 to 90 until reaching apH ~ 50 After fermentation (ca 14 to 18 h) the chubs werecold-showered to an internal temperature of ~80degF (27degC) in thesmokehouse and then dried at 55degF (13degC) and 65 RH [wet bulb50degF (10degC) dry bulb 55degF (13degC)] to a moistureprotein (MPr)ratio of 16 1 (ca 15 to 21 days) in an environmentally controlledchamber (Biotron facility University of Wisconsin Madison WI)In addition to the modified method 6 (27) process just describedfor some experiments chubs were heated to internal temperatures of14SOF (63degC) instantaneous or 128degF (53degC) for 60 minutes andcold-showered before drying Three trials were conducted for eachof the three processes evaluated (i) modified method 6 (ii)modified method 6 with a postfermentation heating of chubs to aninternal temperature of 145degF (63degC) and (iii) modified method 6with a postfermentation heating of chubs to an internal temperatureof 128degF (53degC) for 60 min

Microbiological analyses of pepperoni chubsPepperoni (two chubs per sampling point) was tested for

viable E coli 0157H7 by direct plating prior to stuffing afterfermentation after cooking (when applicable) and after 4 8 1218 and 21 days of drying When numbers of the pathogendecreased below detection by direct plating laquo 102 CFUg) thepresence or absence of the pathogen was determined by enrich-ment The noninoculated batter was also tested for backgroundlevels of E coli 0157H7 and other non-sorbitol-fermentingbacteria by spread plating onto MSA and for total aerobic bacterialnumbers by spread plating onto Trypticase soy agar (TSA) (Difco)plates Immediately after stuffing the chubs were also tested for

1262 HINKENS FAITH LORANG BAILEY BUEGE KASPAR AND LUCHANSKY

viable pediococci by spread plating onto MRS (Difco) agar platesFor sampling chubs the outside of the chub was cleaned with 70ethanol and the casing was removed with the aid of a sterile scalpelA 25-g portion of the batter or a 25-g cross-section from the middleof a chub was asceptically transferred to a Stomacher bag (SewardMedical London UK) containing 225 ml of 01 peptone and thecontents stomached (model 400 Telanar Co Cincinnati OH) for 2min at room temperature One milliliter of the resulting mixturewas serially diluted in 01 peptone and spread plated ontoduplicate MSA or MRS agar plates Another portion (10 ml) wastransferred to 90 ml of modified EC broth (21) containingnovobiocin (final concentration of 20 figml Sigma Chemical CoSt Louis MO) and incubated with shaking (100 rpm) for 18 to 24 hat 37degC to test for the presence or absence of low levels of viable0157H7 cells Following enrichment the samples were diluted in01 peptone and spread plated in duplicate onto MSA Plateswere incubated at 42degC for 48 h to recover E coli 0157H7 and at37degC for 24 h to recover pediococci before colonies were counted

The serotype of six sorbitol nonfermenting (ie white)colonies from MSA plates was determined using an 0157 latexagglutination test (Unipath Limited Basingstoke England) Sorbi-tol-negative 0157-positive isolates were further confirmed as Ecoli using API 20E biochemical test strips (bioMerieux Vitek IncHazelwood MO) and for the H7 antigen by agglutination (Difco)

Chemical analyses of pepperoni chubsAt each sampling point two control chubs (without E coli

0157H7) were removed for sampling The control chubs wereeither transported on ice directly to the Doskocil Companies(Jefferson WI) for testing or held at -20degC for up to seven daysand then delivered to Doskocil Chemical analyses were performedfollowing Association of Official Analytical Chemists (AOAC)procedures (19) Each chub was tested for fat (AOAC procedure96039) moisture (AOAC procedure 95046) protein (AOACprocedure 92806) and salt (AOAC procedure 93547) The wateractivity (aw) was determined using a Rotronic water activity meter(model DT Huntington NY) The pH and titratable acidity (TA)were determined as previously described (18) using a 25-g sampleof sausage macerated in 100 ml of hot (ca 60degC) dH20 The TAwas expressed as the percent lactic acid

Statistical analysisData were analyzed using the Statistical Analysis System

(SAS Institute Cary NC)

RESULTS

Microbiological testing of raw meat and batterAnalyses of meat blocks before inoculation with the

five-strain cocktail and pediococcal starter culture revealedthat none of7 meat blocks tested contained E coli 0157H7by direct plating (data not shown) In addition to counts ofserotype 0157H7 a portion of these same 7 meat blockswas also tested to determine the total aerobic plate countLevels ranged from lt 102 to 4 X 104 CFUg of meat(average 13 X 104 CFUg) indicating that the raw materi-als were of excellent microbiological quality

Prior to addition of the pediococcal starter culture theantimicrobial capabilities of the spice and cure mixtures onthe five-strain mixture of E coli 0157 H7 was also ascer-tained The data revealed that the spice and cure mixturesused in this study had no significant effect on the numbers of0157H7 (a lt 005 P = 09945) The average count of Ecoli 0157H7 from emulsions prepared from 6 of the 7 meatblocks tested prior to the addition of spice and cure was25 X 107 CFUg of batter (range 11 X 106 to 14 X 108CFUg) The average count from these 6 emulsions testedafter the addition of the spice and cure mixtures and grindingwas 22 X 107 cfu per gram of batter (range 15 X 105 to13 X 108 cfugram)

Fermentation at 96degF (36degC) to pH 50 and drying at55degF (BdegC) and 65 RH to a moistureprotein (MPr)ratio of161

The first manufacture process evaluated was the modi-fied method 6 process for pepperoni (Table 1) wherein thepathogen cell numbers remained essentially the same duringthe 14- to 18-h fermentation Although a 12 log CFUg

TABLE 1 E coli 0157H7 count chemical composition and heating regimen of pepperoni during fermentation at 96degF to pH lt50 anddrying at 55degF and 65 RH to a moisture to protein ratio of 5161

Chemical analysis (value SD 11 = 3) for

Sampling time

E coli0157H7(log CFUg SD) pH TA()a Salt () Moisture () Protein () Fat ()

659 29 472 09 77 03 259 Il 95 00 363 09 4177 060 1613 61 3810 220650 47 464 08 86 04 21212 93 01 383 17 3717 116 1757 49 3973 09261437 473 05 88 04 178 17 91 00 377 12 3273 217 1843 97 4293 100597 38 472 09 88 04 153 17 89 00 390 00 2943 189 1933 82 4500 108569 23 474 09 93 09 135 14 87 02 440 29 2747 197 2033 73 4660 144

Day 0Emulsion 689 86Afterfermen- 682 24 48517 62 08 346 07 96 00 317 09 4883 046 1407 plusmn 29 3113 229tatione

Drying day48121821

a TA titratable acidity as lactic acidb MPr moisture to protein ratioC Smokehouse settings were dry bulb 85degF and wet bulb 80degF during conditioning for I h and dry bulb 96degF and wet bulb 92degF duringfermentation to SpH 50

E COLI 0157H7 IN PEPPERONI 1263

a TA titratable acidity as lactic acidb MPr moisture to protein ratioC No positiveno of chubs tested by enrichmentd Smokehouse settings were dry bulb 85degF and wet bulb 80degF during conditioning for 1 h and dry bulb 96degF and wet bulb 92degF duringfermentation to pH 550

e Postfermentation settings were dry bulb 115degF and wet bulb 104degF for 1 h for cycle 1 dry bulb 120degF and wet bulb 102degF for I h for cycle2 dry bulb 125degF and wet bulb 115degF forI h for cycle 3 dry bulb 135degF and wet bulb 126degF for 1 h for cycle 4 and dry bulb 157degF and wetbulb 148degF until 145degF internal temperature reached for cycle 5

decrease was observed after 21 days of drying the desired5-1og CFUg reduction was not achieved The proximatecomposition of the pepperoni displayed the expected levelsfor pH aw salt protein moisture and fat (Table 1) Inrelated experiments the dextrose concentration was adjustedfrom 063 to 090 to extend the fermentation to 24 handlower the pH to lt47 As in the 14- to 18-h fermentationcounts of the pathogen only decreased ca 05 log CFUgafter fermentation and another 141 log CFUg after 12 daysof drying in the single trial conducted (data not shown)These data revealed that the modified method 6 process forpepperoni was not sufficient to eliminate high levels of

serotype 0157H7 strains of E coli following 14 to 24 h offermentation and subsequent drying

Fermentation at 96degF (36degC) to pH5 50 heating to aninternal temperature of 145degF (63degC) and drying at 55degF(13degC) and 65 RH to a MProf5161

Since the modified method 6 process did not produce a5-D kill as required by the USDAFSIS experiments wereconducted to evaluate postfermentation heating for eliminat-ing the pathogen In the first of two such experiments afterreaching pH 5 50 pathogen numbers remained essentiallythe same as starting levels (Table 2) The smokehouse air

a TA titratable acidity as lactic acidb MPr moisture to protein ratioC No positiveno of chubs tested by enrichmentd Smokehouse settings were dry bulb 85degF and wet bulb 80degF during conditioning for 1 h and dry bulb 96degF and wet bulb 92degF duringfermentation to pH 550

e Postfermentation settings were dry bulb 106degF and wet bulb 98degF for I h for cycle 1 dry bulb 120degF and wet bulb 102degF for 1 h for cycle 2dry bulb 130degF and wet bulb 115degF for 1 h for cycle 3 dry bulb 140degF and wet bulb 123degF for I h for cycle 4 and dry bulb 140degF and wetbulb 130degF until 128degF internal temperature reached and then hold for 60 min for cycle 5

1264 HINKENS FAITH LORANG BAILEY BUEGE KASPAR AND LUCHANSKY

temperature was then raised at a rate of 12 to 15degF (70 toSoC) per h until an internal chub temperature of 145degF(63degC) was reached (Table 2) Heating to 145degF (63degC)reduced counts of the pathogen below detection (gt50-logunit decrease) however it was possible to recover viablecells of serotype 0157 H7 by enrichment after heating andup to S days of drying Nevertheless heating chubs to aninternal temperature of 145degF (63degC) delivered the required5-log unit decrease in counts of the pathogen withoutcausing any visible textural or color changes to the finalproduct (data not shown)

Fermentation at 96degF (36degC) to pHS 50 heating to aninternal temperature of 128degF (53degC) for 60 min anddrying at 55degF (13degC) and 65 RH to a MIProfS161

On the basis of the encouraging results with the 145degF(63degC) heating process a second heating regimen wasevaluated As in the previous experiments fermentationalone had no appreciable effect on the pathogen (09-logcfug decrease) (Table 3) However holding the chubs at aninternal temperature of 12soF (53degC) for 60 min (Table 3)resulted in at least a 5-log CFUg reduction in numbers of0157H7 After heating it was only possible to recover thepathogen by enrichment Also heating for 60 min at 12soF(53degC) did not visibly affect the appearance of the resultingproduct and did not cause greasing out (ie there was noresidual fat layer under the casing) (data not shown)

DISCUSSION

Escherichia coli 0157H7 emerged as a foodbornepathogen of primary health concern in the 1990s During thisperiod serotype 0157H7 strains were incriminated innumerous sporadic cases and several outbreaks involving avariety of foods (6 7-9 22 29) Foods of bovine origin ingeneral and ground beef in particular were most oftenidentified as the food vehicles (16) To better understand theresponse of this bacterium to intrinsic (eg pH) andextrinsic (eg temperature) conditions associated with foodsstudies were conducted using apple cider (6 20) cottagecheese (4) ground beef (13 17 26) mayonnaise andordressings (23 29) and salad vegetables (1) In general thesestudies revealed E coli 0157 H7 is very tolerant to acidAlthough there is significant variation in the degree of acidtolerance from strain to strain most 0157H7 strains aremore acid tolerant than other E coli (3 5 11 20) HoweverE coli 0157H7 does not appear to be particularly heattolerant (13) Although there have been reports on thebehavior of non-0157H7 strains of E coli in fermentedmeats (12 28) with the exception of a previous studyconducted at the Food Research Institute using salami (15)no information has been published on the behavior ofserotype 0157H7 strains in fermented sausages

Several factors associated with pepperoni batter wereevaluated for possible effects on the viability of the serotype0157H7 cocktail For example high levels of indigenousbacteria could have a negative impact on the behavior of Ecoli 0157H7 However the relatively low numbers ofaerobic bacteria in the raw materials suggest that indigenous

bacteria were probably not a significant factor in this studyLikewise microbiological analyses of the batter revealedthat the spice and cure mixtures had no significant effect onthe 0157H7 cocktail Due to the limitations of the facilityhousing the smokehouse it was not possible to evaluate theeffect of smoke in this study Smoking is presumably mosteffective against surface contamination of sausage and assuch was not expected to have an appreciable effect on cellsof E coli within pepperoni chubs The addition of smokecould enhance the kill achieved by fermentation but itwould probably not be sufficient to deliver a 5-D killAmong the variety of other factors that could influence theviability of serotype 0157H7 strains in pepperoni the pHand acid type as well as the time held at a given temperatureprobably present the greatest potential for impacting patho-gen numbers (2) For example E coli 0157H7 is moresensitive to acid conditions (pH 36 to 40) at abuse (15 to20degC or 60 to 77degF) rather than refrigeration (4 to SoC or 39to 46degF) temperatures (6 20) It is also significant thatsurvival of serotype 0157H7 strains can also be affected bythe type of acid In a study by Abdul-Raouf et al (2)evaluating the effect of acidulants on E coli 0157H7 inbeef slurries the order of effectiveness for inhibiting growthwas acetic gt lactic gt citric acid at all pH values andtemperatures tested However Conner and Kotrola (10)reported that organic acids and salts enhanced the survival ofserotype 0157H7 strains during storage at 4degC in asynthetic medium These data underscore the importance ofconducting validation studies in bona fide fermented meatsrather than microbiological media andor beaker sausage tomore precisely establish the viability of E coli 0157H7 fora given formulation and manufacturing process

Although lower pH andor higher fermentation tempera-tures may result in a greater decrease in pathogen numbersit is unlikely that these parameters could be adjustedsufficiently to deliver a 5-D kill in pepperoni More specifi-cally at pH S 43 pepperoni may be too tart for mostconsumers Similarly due to intrinsic limitations of thestarter culture it may not be possible to conduct a fermenta-tion at temperatures much in excess of 120degF (49degC) Forthese reasons it was necessary to evaluate postfermentationheating of chubs as a method to deliver the mandated 5-Dreduction in counts of E coli 0157H7 The first thermalprocess tested 145degF (63degC) internal was selected becauseit meets the requirements for trichinae destruction andbecause it approximates the time temperature and humidityguidelines established by the USDA for cooked andor roastbeef (27) The cooking requirements for cooked andor roastbeef were also approved by the USDA FSIS as an option tocontrol serotype 0157H7 strains of E coli in fermentedsausage (25) The other thermal process tested 12soF(53degC) internal for 60 min was selected because it alsomeets the requirements for trichinae control but moreimportant it is less harmful to the texture and appearance ofpepperoni than the 145degF (63degC) cooking regimen (SteveSeideman personal communication) This approach is alsonot without limitations as some sausage varieties are notamenable to postfermentation heating due to possible unde-

E COLI 0157H7 IN PEPPERONI 1265

sirable textural effects on the product andor some manufac-turers are not capable of including such a heating step due tothe volume of product produced or the cost of the attendantequipment and facilities From a practical standpoint heat-ing may cause greasing out and may yield a product moreprone to cupping when used as a pizza topping (SteveSeideman personal communication) Further studies may bewarranted to identify heating regimens that are sufficient toeliminate the pathogen without producing untoward effectson the finished product and that are cost effective

The present study evaluated the modified method 6pepperoni process as well as postfermentation heating toinactivate E coli 0157H7 A comprehensive sampling andscreening of raw materials to account for low levels of thispathogen in raw materials as well as adherence to goodmanufacturing procedures would also contribute to thesafety of fermented meat products As another considerationthe sampling plan for raw materials that will undergosubsequent heat processing such as pepperoni used as pizzatopping is less stringent (ie 15 samples per lot comparedto 30 samples per lot [25 J) because the ensuing cooking ispresumably sufficient to eliminate the pathogen Perhaps amore serious threat to consumer safety would result fromcontaminated pepperoni used for sandwiches or salad barsFor these uses it is necessary to include a postfermentationheating step sufficient to deliver the 5-D kill Thus it wouldalso be informative to substantiate the fate of the pathogenafter frozen or refrigerated storage of pepperoni targeted foruse in sandwiches on pizza or for salad bars

The results of the present study revealed that thenonthermal process (modified method 6) for pepperonimanufacture was not sufficient to effect a 5-10g unit reduc-tion in counts of E coli 0157H7 The observed lt2-10g unitdecrease in counts of the pathogen following fermentation ofpepperoni at 96degF (36degC) to pH ~ 50 and drying was inagreement with a previous report of a 1- to 2-10g unitdecrease in counts of the pathogen under similar processconditions for salami (ie fermentation at 96degF (36degC) topH 48 drying to a MPr of ~191 and storage at 4degC for 2weeks) (15) In the present study postfermentation heatingachieved a 5-D kill of this pathogen Although both heatingregimens evaluated were effective for eliminating E coliOI57H7 cooking at l28degF (53degC) for 60 min was lessharmful to the product than cooking to an internal tempera-ture of 145degF (63degC) These data represent the first of severalreports expected to be published on the fate of E coli0157H7 in fermented meats by our group and by otherinvestigators In addition to a thermal processing step weare evaluating the eff~ct of fermentation temperature andpH a 3- to 7-day postfermentation storage at low pH andvarious heating regimens for eliminating this pathogen inready-to-eat fermented sausage It is possible that slightincreases in the fermentation temperature andor extendedstorage at low pH would result in a greater reduction incounts of the pathogen and further reduce the time andtemperature required for postfermentation heating A compre-hensive effort to screen raw materials for the pathogen tofollow good manufacturing practices and to monitor fin-ished product would significantly reduce the likelihood of

illness due to E coli 0157H7 in fermented sausage In theinterim the results of the present study will assist manufac-turers in prioritizing options for validating the safety of theirrespective manufacturing processes

ACKNOWLEDGMENTS

The technical assistance of Patrick Krakar Tamara Dodge and KarenArnold of the Food Research Institute and Rong Chen and Jane Wagner ofDoskocil Foods are gratefully acknowledged Our appreciation is alsoextended to Chuck Baum (OW-Madison Biotron) John Cerveny (OscarMayer) Steven Devcich (TeePak) Gerry Durnell (National Association ofPizza Operators) Larry Hand (Diversitech) Steve Lary (Alkar) and SteveSeideman (Doskocil) for supplies service equipment andor consultationsWe also thank the following individuals for supplying cultures Jay HLewis (Washington State Public Health Lab) for strain F-90 Mary EProctor (Wisconsin Bureau of Public Health) for strain SLH21788 KathyGlass (Food Research Institute) for strain C7927 and Joy G Wells andTimothy J Barrett (Centers for Disease Control and Prevention) for strainC9490

This work was supported by a gift from Doskocil Companies Inc andthe College of Agricultural and Life Sciences of the University ofWisconsin-Madison

REFERENCES

1 Abdul-Raouf U M L R Beuchat and M S Ammar 1993 Survivaland growth of Escherichia coli 0157H7 on salad vegetables ApplEnviron Microbiol 59 1999-2006

2 Abdul-Raouf U M L R Beuchat and M S Ammar 1993 Survivaland growth of Escherichia coli 0157H7 in ground roasted beef asaffected by pH acidulants and temperature Appl Environ Micro-bioI 592364-2368

3 Arnold K W and C W Kaspar 1995 Starvation- and stationary-phase-induced acid tolerance in Escherichia coli 0157H7 ApplEnviron Microbiol 612037-2039

4 Arocha M M M McVey S D Loder J H Rupnow and L BBullerman 1992 Behavior of hemorrhagic Escherichia coli 0157H7during the manufacture of cottage cheese J Food Prot 55379-381

5 Benjamin M M andA R Datta 1995 Acid tolerance of enterohem-orrhagic Escherichia coli Appl Environ Microbiol 61 1669-1672

6 Besser R E S M Lett J T Weber M P Doyle T J Barrett J GWells and P M Griffin 1993 An outbreak of diarrhea and hemolyticuremic syndrome from Escherichia coli 0157H7 in fresh-pressedapple cider JAMA 2692217-2220

7 Centers for Disease Control and Prevention 1993 Update multi stateoutbreak of Escherichia coli 0157H7 infections from hamburgers-western United States 1992-1993 Morbid Mortal Weekly Rep42258-263

8 Centers for Disease Control and Prevention 1995 Escherichia coli0157H7 outbreak linked to commercially distributed dry-curedsalami-Washington and California 1994 Morbid Mortal WeeklyRep 44157-160

9 Centers for Disease Control and Prevention 1995 Communityoutbreak of hemolytic uremic syndrome attributable to Escherichiacoli 0111 NM-South Australia 1995 Morbid Mortal Weekly Rep44550-551 557

10 Conner D E and J S~ Kotrola 1995 Growth and survival ofEscherichia coli 0157H7 under acidic conditions Appl EnvironMicrobiol 61382-385

l1 Cutter C N and G R Siragusa 1994 Efficacy of organic acidsagainst Escherichia coli 0157H7 attached to beef carcass tissue usinga pilot scale model carcass washer J Food Prot 5797-103

12 Darmadjii P M Izumimoto T Miyamoto and K Katoaka 1990Lactic fermentation on preservative qualities of Dendeng giling JFood Sci 55(6) 1523-1527

13 Doyle M P and J L Schoeni 1984 Survival and growthcharacteristics of Escherichia coli associated with hemorrhagiccolitis Appl Environ Microbiol 48855-856

14 Faith N G J Hinkens T D Lorang P Bailey D Buege C WKaspar and J B Luchansky 1996 Validation of pepperoni processesfor control of Escherichia coli 0157H7 abstr 191 p 80 Progr

1262 HINKENS FAITH LORANG BAILEY BUEGE KASPAR AND LUCHANSKY

viable pediococci by spread plating onto MRS (Difco) agar platesFor sampling chubs the outside of the chub was cleaned with 70ethanol and the casing was removed with the aid of a sterile scalpelA 25-g portion of the batter or a 25-g cross-section from the middleof a chub was asceptically transferred to a Stomacher bag (SewardMedical London UK) containing 225 ml of 01 peptone and thecontents stomached (model 400 Telanar Co Cincinnati OH) for 2min at room temperature One milliliter of the resulting mixturewas serially diluted in 01 peptone and spread plated ontoduplicate MSA or MRS agar plates Another portion (10 ml) wastransferred to 90 ml of modified EC broth (21) containingnovobiocin (final concentration of 20 figml Sigma Chemical CoSt Louis MO) and incubated with shaking (100 rpm) for 18 to 24 hat 37degC to test for the presence or absence of low levels of viable0157H7 cells Following enrichment the samples were diluted in01 peptone and spread plated in duplicate onto MSA Plateswere incubated at 42degC for 48 h to recover E coli 0157H7 and at37degC for 24 h to recover pediococci before colonies were counted

The serotype of six sorbitol nonfermenting (ie white)colonies from MSA plates was determined using an 0157 latexagglutination test (Unipath Limited Basingstoke England) Sorbi-tol-negative 0157-positive isolates were further confirmed as Ecoli using API 20E biochemical test strips (bioMerieux Vitek IncHazelwood MO) and for the H7 antigen by agglutination (Difco)

Chemical analyses of pepperoni chubsAt each sampling point two control chubs (without E coli

0157H7) were removed for sampling The control chubs wereeither transported on ice directly to the Doskocil Companies(Jefferson WI) for testing or held at -20degC for up to seven daysand then delivered to Doskocil Chemical analyses were performedfollowing Association of Official Analytical Chemists (AOAC)procedures (19) Each chub was tested for fat (AOAC procedure96039) moisture (AOAC procedure 95046) protein (AOACprocedure 92806) and salt (AOAC procedure 93547) The wateractivity (aw) was determined using a Rotronic water activity meter(model DT Huntington NY) The pH and titratable acidity (TA)were determined as previously described (18) using a 25-g sampleof sausage macerated in 100 ml of hot (ca 60degC) dH20 The TAwas expressed as the percent lactic acid

Statistical analysisData were analyzed using the Statistical Analysis System

(SAS Institute Cary NC)

RESULTS

Microbiological testing of raw meat and batterAnalyses of meat blocks before inoculation with the

five-strain cocktail and pediococcal starter culture revealedthat none of7 meat blocks tested contained E coli 0157H7by direct plating (data not shown) In addition to counts ofserotype 0157H7 a portion of these same 7 meat blockswas also tested to determine the total aerobic plate countLevels ranged from lt 102 to 4 X 104 CFUg of meat(average 13 X 104 CFUg) indicating that the raw materi-als were of excellent microbiological quality

Prior to addition of the pediococcal starter culture theantimicrobial capabilities of the spice and cure mixtures onthe five-strain mixture of E coli 0157 H7 was also ascer-tained The data revealed that the spice and cure mixturesused in this study had no significant effect on the numbers of0157H7 (a lt 005 P = 09945) The average count of Ecoli 0157H7 from emulsions prepared from 6 of the 7 meatblocks tested prior to the addition of spice and cure was25 X 107 CFUg of batter (range 11 X 106 to 14 X 108CFUg) The average count from these 6 emulsions testedafter the addition of the spice and cure mixtures and grindingwas 22 X 107 cfu per gram of batter (range 15 X 105 to13 X 108 cfugram)

Fermentation at 96degF (36degC) to pH 50 and drying at55degF (BdegC) and 65 RH to a moistureprotein (MPr)ratio of161

The first manufacture process evaluated was the modi-fied method 6 process for pepperoni (Table 1) wherein thepathogen cell numbers remained essentially the same duringthe 14- to 18-h fermentation Although a 12 log CFUg

TABLE 1 E coli 0157H7 count chemical composition and heating regimen of pepperoni during fermentation at 96degF to pH lt50 anddrying at 55degF and 65 RH to a moisture to protein ratio of 5161

Chemical analysis (value SD 11 = 3) for

Sampling time

E coli0157H7(log CFUg SD) pH TA()a Salt () Moisture () Protein () Fat ()

659 29 472 09 77 03 259 Il 95 00 363 09 4177 060 1613 61 3810 220650 47 464 08 86 04 21212 93 01 383 17 3717 116 1757 49 3973 09261437 473 05 88 04 178 17 91 00 377 12 3273 217 1843 97 4293 100597 38 472 09 88 04 153 17 89 00 390 00 2943 189 1933 82 4500 108569 23 474 09 93 09 135 14 87 02 440 29 2747 197 2033 73 4660 144

Day 0Emulsion 689 86Afterfermen- 682 24 48517 62 08 346 07 96 00 317 09 4883 046 1407 plusmn 29 3113 229tatione

Drying day48121821

a TA titratable acidity as lactic acidb MPr moisture to protein ratioC Smokehouse settings were dry bulb 85degF and wet bulb 80degF during conditioning for I h and dry bulb 96degF and wet bulb 92degF duringfermentation to SpH 50

E COLI 0157H7 IN PEPPERONI 1263

a TA titratable acidity as lactic acidb MPr moisture to protein ratioC No positiveno of chubs tested by enrichmentd Smokehouse settings were dry bulb 85degF and wet bulb 80degF during conditioning for 1 h and dry bulb 96degF and wet bulb 92degF duringfermentation to pH 550

e Postfermentation settings were dry bulb 115degF and wet bulb 104degF for 1 h for cycle 1 dry bulb 120degF and wet bulb 102degF for I h for cycle2 dry bulb 125degF and wet bulb 115degF forI h for cycle 3 dry bulb 135degF and wet bulb 126degF for 1 h for cycle 4 and dry bulb 157degF and wetbulb 148degF until 145degF internal temperature reached for cycle 5

decrease was observed after 21 days of drying the desired5-1og CFUg reduction was not achieved The proximatecomposition of the pepperoni displayed the expected levelsfor pH aw salt protein moisture and fat (Table 1) Inrelated experiments the dextrose concentration was adjustedfrom 063 to 090 to extend the fermentation to 24 handlower the pH to lt47 As in the 14- to 18-h fermentationcounts of the pathogen only decreased ca 05 log CFUgafter fermentation and another 141 log CFUg after 12 daysof drying in the single trial conducted (data not shown)These data revealed that the modified method 6 process forpepperoni was not sufficient to eliminate high levels of

serotype 0157H7 strains of E coli following 14 to 24 h offermentation and subsequent drying

Fermentation at 96degF (36degC) to pH5 50 heating to aninternal temperature of 145degF (63degC) and drying at 55degF(13degC) and 65 RH to a MProf5161

Since the modified method 6 process did not produce a5-D kill as required by the USDAFSIS experiments wereconducted to evaluate postfermentation heating for eliminat-ing the pathogen In the first of two such experiments afterreaching pH 5 50 pathogen numbers remained essentiallythe same as starting levels (Table 2) The smokehouse air

a TA titratable acidity as lactic acidb MPr moisture to protein ratioC No positiveno of chubs tested by enrichmentd Smokehouse settings were dry bulb 85degF and wet bulb 80degF during conditioning for 1 h and dry bulb 96degF and wet bulb 92degF duringfermentation to pH 550

e Postfermentation settings were dry bulb 106degF and wet bulb 98degF for I h for cycle 1 dry bulb 120degF and wet bulb 102degF for 1 h for cycle 2dry bulb 130degF and wet bulb 115degF for 1 h for cycle 3 dry bulb 140degF and wet bulb 123degF for I h for cycle 4 and dry bulb 140degF and wetbulb 130degF until 128degF internal temperature reached and then hold for 60 min for cycle 5

1264 HINKENS FAITH LORANG BAILEY BUEGE KASPAR AND LUCHANSKY

temperature was then raised at a rate of 12 to 15degF (70 toSoC) per h until an internal chub temperature of 145degF(63degC) was reached (Table 2) Heating to 145degF (63degC)reduced counts of the pathogen below detection (gt50-logunit decrease) however it was possible to recover viablecells of serotype 0157 H7 by enrichment after heating andup to S days of drying Nevertheless heating chubs to aninternal temperature of 145degF (63degC) delivered the required5-log unit decrease in counts of the pathogen withoutcausing any visible textural or color changes to the finalproduct (data not shown)

Fermentation at 96degF (36degC) to pHS 50 heating to aninternal temperature of 128degF (53degC) for 60 min anddrying at 55degF (13degC) and 65 RH to a MIProfS161

On the basis of the encouraging results with the 145degF(63degC) heating process a second heating regimen wasevaluated As in the previous experiments fermentationalone had no appreciable effect on the pathogen (09-logcfug decrease) (Table 3) However holding the chubs at aninternal temperature of 12soF (53degC) for 60 min (Table 3)resulted in at least a 5-log CFUg reduction in numbers of0157H7 After heating it was only possible to recover thepathogen by enrichment Also heating for 60 min at 12soF(53degC) did not visibly affect the appearance of the resultingproduct and did not cause greasing out (ie there was noresidual fat layer under the casing) (data not shown)

DISCUSSION

Escherichia coli 0157H7 emerged as a foodbornepathogen of primary health concern in the 1990s During thisperiod serotype 0157H7 strains were incriminated innumerous sporadic cases and several outbreaks involving avariety of foods (6 7-9 22 29) Foods of bovine origin ingeneral and ground beef in particular were most oftenidentified as the food vehicles (16) To better understand theresponse of this bacterium to intrinsic (eg pH) andextrinsic (eg temperature) conditions associated with foodsstudies were conducted using apple cider (6 20) cottagecheese (4) ground beef (13 17 26) mayonnaise andordressings (23 29) and salad vegetables (1) In general thesestudies revealed E coli 0157 H7 is very tolerant to acidAlthough there is significant variation in the degree of acidtolerance from strain to strain most 0157H7 strains aremore acid tolerant than other E coli (3 5 11 20) HoweverE coli 0157H7 does not appear to be particularly heattolerant (13) Although there have been reports on thebehavior of non-0157H7 strains of E coli in fermentedmeats (12 28) with the exception of a previous studyconducted at the Food Research Institute using salami (15)no information has been published on the behavior ofserotype 0157H7 strains in fermented sausages

Several factors associated with pepperoni batter wereevaluated for possible effects on the viability of the serotype0157H7 cocktail For example high levels of indigenousbacteria could have a negative impact on the behavior of Ecoli 0157H7 However the relatively low numbers ofaerobic bacteria in the raw materials suggest that indigenous

bacteria were probably not a significant factor in this studyLikewise microbiological analyses of the batter revealedthat the spice and cure mixtures had no significant effect onthe 0157H7 cocktail Due to the limitations of the facilityhousing the smokehouse it was not possible to evaluate theeffect of smoke in this study Smoking is presumably mosteffective against surface contamination of sausage and assuch was not expected to have an appreciable effect on cellsof E coli within pepperoni chubs The addition of smokecould enhance the kill achieved by fermentation but itwould probably not be sufficient to deliver a 5-D killAmong the variety of other factors that could influence theviability of serotype 0157H7 strains in pepperoni the pHand acid type as well as the time held at a given temperatureprobably present the greatest potential for impacting patho-gen numbers (2) For example E coli 0157H7 is moresensitive to acid conditions (pH 36 to 40) at abuse (15 to20degC or 60 to 77degF) rather than refrigeration (4 to SoC or 39to 46degF) temperatures (6 20) It is also significant thatsurvival of serotype 0157H7 strains can also be affected bythe type of acid In a study by Abdul-Raouf et al (2)evaluating the effect of acidulants on E coli 0157H7 inbeef slurries the order of effectiveness for inhibiting growthwas acetic gt lactic gt citric acid at all pH values andtemperatures tested However Conner and Kotrola (10)reported that organic acids and salts enhanced the survival ofserotype 0157H7 strains during storage at 4degC in asynthetic medium These data underscore the importance ofconducting validation studies in bona fide fermented meatsrather than microbiological media andor beaker sausage tomore precisely establish the viability of E coli 0157H7 fora given formulation and manufacturing process

Although lower pH andor higher fermentation tempera-tures may result in a greater decrease in pathogen numbersit is unlikely that these parameters could be adjustedsufficiently to deliver a 5-D kill in pepperoni More specifi-cally at pH S 43 pepperoni may be too tart for mostconsumers Similarly due to intrinsic limitations of thestarter culture it may not be possible to conduct a fermenta-tion at temperatures much in excess of 120degF (49degC) Forthese reasons it was necessary to evaluate postfermentationheating of chubs as a method to deliver the mandated 5-Dreduction in counts of E coli 0157H7 The first thermalprocess tested 145degF (63degC) internal was selected becauseit meets the requirements for trichinae destruction andbecause it approximates the time temperature and humidityguidelines established by the USDA for cooked andor roastbeef (27) The cooking requirements for cooked andor roastbeef were also approved by the USDA FSIS as an option tocontrol serotype 0157H7 strains of E coli in fermentedsausage (25) The other thermal process tested 12soF(53degC) internal for 60 min was selected because it alsomeets the requirements for trichinae control but moreimportant it is less harmful to the texture and appearance ofpepperoni than the 145degF (63degC) cooking regimen (SteveSeideman personal communication) This approach is alsonot without limitations as some sausage varieties are notamenable to postfermentation heating due to possible unde-

E COLI 0157H7 IN PEPPERONI 1265

sirable textural effects on the product andor some manufac-turers are not capable of including such a heating step due tothe volume of product produced or the cost of the attendantequipment and facilities From a practical standpoint heat-ing may cause greasing out and may yield a product moreprone to cupping when used as a pizza topping (SteveSeideman personal communication) Further studies may bewarranted to identify heating regimens that are sufficient toeliminate the pathogen without producing untoward effectson the finished product and that are cost effective

The present study evaluated the modified method 6pepperoni process as well as postfermentation heating toinactivate E coli 0157H7 A comprehensive sampling andscreening of raw materials to account for low levels of thispathogen in raw materials as well as adherence to goodmanufacturing procedures would also contribute to thesafety of fermented meat products As another considerationthe sampling plan for raw materials that will undergosubsequent heat processing such as pepperoni used as pizzatopping is less stringent (ie 15 samples per lot comparedto 30 samples per lot [25 J) because the ensuing cooking ispresumably sufficient to eliminate the pathogen Perhaps amore serious threat to consumer safety would result fromcontaminated pepperoni used for sandwiches or salad barsFor these uses it is necessary to include a postfermentationheating step sufficient to deliver the 5-D kill Thus it wouldalso be informative to substantiate the fate of the pathogenafter frozen or refrigerated storage of pepperoni targeted foruse in sandwiches on pizza or for salad bars

The results of the present study revealed that thenonthermal process (modified method 6) for pepperonimanufacture was not sufficient to effect a 5-10g unit reduc-tion in counts of E coli 0157H7 The observed lt2-10g unitdecrease in counts of the pathogen following fermentation ofpepperoni at 96degF (36degC) to pH ~ 50 and drying was inagreement with a previous report of a 1- to 2-10g unitdecrease in counts of the pathogen under similar processconditions for salami (ie fermentation at 96degF (36degC) topH 48 drying to a MPr of ~191 and storage at 4degC for 2weeks) (15) In the present study postfermentation heatingachieved a 5-D kill of this pathogen Although both heatingregimens evaluated were effective for eliminating E coliOI57H7 cooking at l28degF (53degC) for 60 min was lessharmful to the product than cooking to an internal tempera-ture of 145degF (63degC) These data represent the first of severalreports expected to be published on the fate of E coli0157H7 in fermented meats by our group and by otherinvestigators In addition to a thermal processing step weare evaluating the eff~ct of fermentation temperature andpH a 3- to 7-day postfermentation storage at low pH andvarious heating regimens for eliminating this pathogen inready-to-eat fermented sausage It is possible that slightincreases in the fermentation temperature andor extendedstorage at low pH would result in a greater reduction incounts of the pathogen and further reduce the time andtemperature required for postfermentation heating A compre-hensive effort to screen raw materials for the pathogen tofollow good manufacturing practices and to monitor fin-ished product would significantly reduce the likelihood of

illness due to E coli 0157H7 in fermented sausage In theinterim the results of the present study will assist manufac-turers in prioritizing options for validating the safety of theirrespective manufacturing processes

ACKNOWLEDGMENTS

The technical assistance of Patrick Krakar Tamara Dodge and KarenArnold of the Food Research Institute and Rong Chen and Jane Wagner ofDoskocil Foods are gratefully acknowledged Our appreciation is alsoextended to Chuck Baum (OW-Madison Biotron) John Cerveny (OscarMayer) Steven Devcich (TeePak) Gerry Durnell (National Association ofPizza Operators) Larry Hand (Diversitech) Steve Lary (Alkar) and SteveSeideman (Doskocil) for supplies service equipment andor consultationsWe also thank the following individuals for supplying cultures Jay HLewis (Washington State Public Health Lab) for strain F-90 Mary EProctor (Wisconsin Bureau of Public Health) for strain SLH21788 KathyGlass (Food Research Institute) for strain C7927 and Joy G Wells andTimothy J Barrett (Centers for Disease Control and Prevention) for strainC9490

This work was supported by a gift from Doskocil Companies Inc andthe College of Agricultural and Life Sciences of the University ofWisconsin-Madison

REFERENCES

1 Abdul-Raouf U M L R Beuchat and M S Ammar 1993 Survivaland growth of Escherichia coli 0157H7 on salad vegetables ApplEnviron Microbiol 59 1999-2006

2 Abdul-Raouf U M L R Beuchat and M S Ammar 1993 Survivaland growth of Escherichia coli 0157H7 in ground roasted beef asaffected by pH acidulants and temperature Appl Environ Micro-bioI 592364-2368

3 Arnold K W and C W Kaspar 1995 Starvation- and stationary-phase-induced acid tolerance in Escherichia coli 0157H7 ApplEnviron Microbiol 612037-2039

4 Arocha M M M McVey S D Loder J H Rupnow and L BBullerman 1992 Behavior of hemorrhagic Escherichia coli 0157H7during the manufacture of cottage cheese J Food Prot 55379-381

5 Benjamin M M andA R Datta 1995 Acid tolerance of enterohem-orrhagic Escherichia coli Appl Environ Microbiol 61 1669-1672

6 Besser R E S M Lett J T Weber M P Doyle T J Barrett J GWells and P M Griffin 1993 An outbreak of diarrhea and hemolyticuremic syndrome from Escherichia coli 0157H7 in fresh-pressedapple cider JAMA 2692217-2220

7 Centers for Disease Control and Prevention 1993 Update multi stateoutbreak of Escherichia coli 0157H7 infections from hamburgers-western United States 1992-1993 Morbid Mortal Weekly Rep42258-263

8 Centers for Disease Control and Prevention 1995 Escherichia coli0157H7 outbreak linked to commercially distributed dry-curedsalami-Washington and California 1994 Morbid Mortal WeeklyRep 44157-160

9 Centers for Disease Control and Prevention 1995 Communityoutbreak of hemolytic uremic syndrome attributable to Escherichiacoli 0111 NM-South Australia 1995 Morbid Mortal Weekly Rep44550-551 557

10 Conner D E and J S~ Kotrola 1995 Growth and survival ofEscherichia coli 0157H7 under acidic conditions Appl EnvironMicrobiol 61382-385

l1 Cutter C N and G R Siragusa 1994 Efficacy of organic acidsagainst Escherichia coli 0157H7 attached to beef carcass tissue usinga pilot scale model carcass washer J Food Prot 5797-103

12 Darmadjii P M Izumimoto T Miyamoto and K Katoaka 1990Lactic fermentation on preservative qualities of Dendeng giling JFood Sci 55(6) 1523-1527

13 Doyle M P and J L Schoeni 1984 Survival and growthcharacteristics of Escherichia coli associated with hemorrhagiccolitis Appl Environ Microbiol 48855-856

14 Faith N G J Hinkens T D Lorang P Bailey D Buege C WKaspar and J B Luchansky 1996 Validation of pepperoni processesfor control of Escherichia coli 0157H7 abstr 191 p 80 Progr

E COLI 0157H7 IN PEPPERONI 1263

a TA titratable acidity as lactic acidb MPr moisture to protein ratioC No positiveno of chubs tested by enrichmentd Smokehouse settings were dry bulb 85degF and wet bulb 80degF during conditioning for 1 h and dry bulb 96degF and wet bulb 92degF duringfermentation to pH 550

e Postfermentation settings were dry bulb 115degF and wet bulb 104degF for 1 h for cycle 1 dry bulb 120degF and wet bulb 102degF for I h for cycle2 dry bulb 125degF and wet bulb 115degF forI h for cycle 3 dry bulb 135degF and wet bulb 126degF for 1 h for cycle 4 and dry bulb 157degF and wetbulb 148degF until 145degF internal temperature reached for cycle 5

decrease was observed after 21 days of drying the desired5-1og CFUg reduction was not achieved The proximatecomposition of the pepperoni displayed the expected levelsfor pH aw salt protein moisture and fat (Table 1) Inrelated experiments the dextrose concentration was adjustedfrom 063 to 090 to extend the fermentation to 24 handlower the pH to lt47 As in the 14- to 18-h fermentationcounts of the pathogen only decreased ca 05 log CFUgafter fermentation and another 141 log CFUg after 12 daysof drying in the single trial conducted (data not shown)These data revealed that the modified method 6 process forpepperoni was not sufficient to eliminate high levels of

serotype 0157H7 strains of E coli following 14 to 24 h offermentation and subsequent drying

Fermentation at 96degF (36degC) to pH5 50 heating to aninternal temperature of 145degF (63degC) and drying at 55degF(13degC) and 65 RH to a MProf5161

Since the modified method 6 process did not produce a5-D kill as required by the USDAFSIS experiments wereconducted to evaluate postfermentation heating for eliminat-ing the pathogen In the first of two such experiments afterreaching pH 5 50 pathogen numbers remained essentiallythe same as starting levels (Table 2) The smokehouse air

a TA titratable acidity as lactic acidb MPr moisture to protein ratioC No positiveno of chubs tested by enrichmentd Smokehouse settings were dry bulb 85degF and wet bulb 80degF during conditioning for 1 h and dry bulb 96degF and wet bulb 92degF duringfermentation to pH 550

e Postfermentation settings were dry bulb 106degF and wet bulb 98degF for I h for cycle 1 dry bulb 120degF and wet bulb 102degF for 1 h for cycle 2dry bulb 130degF and wet bulb 115degF for 1 h for cycle 3 dry bulb 140degF and wet bulb 123degF for I h for cycle 4 and dry bulb 140degF and wetbulb 130degF until 128degF internal temperature reached and then hold for 60 min for cycle 5

1264 HINKENS FAITH LORANG BAILEY BUEGE KASPAR AND LUCHANSKY

temperature was then raised at a rate of 12 to 15degF (70 toSoC) per h until an internal chub temperature of 145degF(63degC) was reached (Table 2) Heating to 145degF (63degC)reduced counts of the pathogen below detection (gt50-logunit decrease) however it was possible to recover viablecells of serotype 0157 H7 by enrichment after heating andup to S days of drying Nevertheless heating chubs to aninternal temperature of 145degF (63degC) delivered the required5-log unit decrease in counts of the pathogen withoutcausing any visible textural or color changes to the finalproduct (data not shown)

Fermentation at 96degF (36degC) to pHS 50 heating to aninternal temperature of 128degF (53degC) for 60 min anddrying at 55degF (13degC) and 65 RH to a MIProfS161

On the basis of the encouraging results with the 145degF(63degC) heating process a second heating regimen wasevaluated As in the previous experiments fermentationalone had no appreciable effect on the pathogen (09-logcfug decrease) (Table 3) However holding the chubs at aninternal temperature of 12soF (53degC) for 60 min (Table 3)resulted in at least a 5-log CFUg reduction in numbers of0157H7 After heating it was only possible to recover thepathogen by enrichment Also heating for 60 min at 12soF(53degC) did not visibly affect the appearance of the resultingproduct and did not cause greasing out (ie there was noresidual fat layer under the casing) (data not shown)

DISCUSSION

Escherichia coli 0157H7 emerged as a foodbornepathogen of primary health concern in the 1990s During thisperiod serotype 0157H7 strains were incriminated innumerous sporadic cases and several outbreaks involving avariety of foods (6 7-9 22 29) Foods of bovine origin ingeneral and ground beef in particular were most oftenidentified as the food vehicles (16) To better understand theresponse of this bacterium to intrinsic (eg pH) andextrinsic (eg temperature) conditions associated with foodsstudies were conducted using apple cider (6 20) cottagecheese (4) ground beef (13 17 26) mayonnaise andordressings (23 29) and salad vegetables (1) In general thesestudies revealed E coli 0157 H7 is very tolerant to acidAlthough there is significant variation in the degree of acidtolerance from strain to strain most 0157H7 strains aremore acid tolerant than other E coli (3 5 11 20) HoweverE coli 0157H7 does not appear to be particularly heattolerant (13) Although there have been reports on thebehavior of non-0157H7 strains of E coli in fermentedmeats (12 28) with the exception of a previous studyconducted at the Food Research Institute using salami (15)no information has been published on the behavior ofserotype 0157H7 strains in fermented sausages

Several factors associated with pepperoni batter wereevaluated for possible effects on the viability of the serotype0157H7 cocktail For example high levels of indigenousbacteria could have a negative impact on the behavior of Ecoli 0157H7 However the relatively low numbers ofaerobic bacteria in the raw materials suggest that indigenous

bacteria were probably not a significant factor in this studyLikewise microbiological analyses of the batter revealedthat the spice and cure mixtures had no significant effect onthe 0157H7 cocktail Due to the limitations of the facilityhousing the smokehouse it was not possible to evaluate theeffect of smoke in this study Smoking is presumably mosteffective against surface contamination of sausage and assuch was not expected to have an appreciable effect on cellsof E coli within pepperoni chubs The addition of smokecould enhance the kill achieved by fermentation but itwould probably not be sufficient to deliver a 5-D killAmong the variety of other factors that could influence theviability of serotype 0157H7 strains in pepperoni the pHand acid type as well as the time held at a given temperatureprobably present the greatest potential for impacting patho-gen numbers (2) For example E coli 0157H7 is moresensitive to acid conditions (pH 36 to 40) at abuse (15 to20degC or 60 to 77degF) rather than refrigeration (4 to SoC or 39to 46degF) temperatures (6 20) It is also significant thatsurvival of serotype 0157H7 strains can also be affected bythe type of acid In a study by Abdul-Raouf et al (2)evaluating the effect of acidulants on E coli 0157H7 inbeef slurries the order of effectiveness for inhibiting growthwas acetic gt lactic gt citric acid at all pH values andtemperatures tested However Conner and Kotrola (10)reported that organic acids and salts enhanced the survival ofserotype 0157H7 strains during storage at 4degC in asynthetic medium These data underscore the importance ofconducting validation studies in bona fide fermented meatsrather than microbiological media andor beaker sausage tomore precisely establish the viability of E coli 0157H7 fora given formulation and manufacturing process

Although lower pH andor higher fermentation tempera-tures may result in a greater decrease in pathogen numbersit is unlikely that these parameters could be adjustedsufficiently to deliver a 5-D kill in pepperoni More specifi-cally at pH S 43 pepperoni may be too tart for mostconsumers Similarly due to intrinsic limitations of thestarter culture it may not be possible to conduct a fermenta-tion at temperatures much in excess of 120degF (49degC) Forthese reasons it was necessary to evaluate postfermentationheating of chubs as a method to deliver the mandated 5-Dreduction in counts of E coli 0157H7 The first thermalprocess tested 145degF (63degC) internal was selected becauseit meets the requirements for trichinae destruction andbecause it approximates the time temperature and humidityguidelines established by the USDA for cooked andor roastbeef (27) The cooking requirements for cooked andor roastbeef were also approved by the USDA FSIS as an option tocontrol serotype 0157H7 strains of E coli in fermentedsausage (25) The other thermal process tested 12soF(53degC) internal for 60 min was selected because it alsomeets the requirements for trichinae control but moreimportant it is less harmful to the texture and appearance ofpepperoni than the 145degF (63degC) cooking regimen (SteveSeideman personal communication) This approach is alsonot without limitations as some sausage varieties are notamenable to postfermentation heating due to possible unde-

E COLI 0157H7 IN PEPPERONI 1265

sirable textural effects on the product andor some manufac-turers are not capable of including such a heating step due tothe volume of product produced or the cost of the attendantequipment and facilities From a practical standpoint heat-ing may cause greasing out and may yield a product moreprone to cupping when used as a pizza topping (SteveSeideman personal communication) Further studies may bewarranted to identify heating regimens that are sufficient toeliminate the pathogen without producing untoward effectson the finished product and that are cost effective

The present study evaluated the modified method 6pepperoni process as well as postfermentation heating toinactivate E coli 0157H7 A comprehensive sampling andscreening of raw materials to account for low levels of thispathogen in raw materials as well as adherence to goodmanufacturing procedures would also contribute to thesafety of fermented meat products As another considerationthe sampling plan for raw materials that will undergosubsequent heat processing such as pepperoni used as pizzatopping is less stringent (ie 15 samples per lot comparedto 30 samples per lot [25 J) because the ensuing cooking ispresumably sufficient to eliminate the pathogen Perhaps amore serious threat to consumer safety would result fromcontaminated pepperoni used for sandwiches or salad barsFor these uses it is necessary to include a postfermentationheating step sufficient to deliver the 5-D kill Thus it wouldalso be informative to substantiate the fate of the pathogenafter frozen or refrigerated storage of pepperoni targeted foruse in sandwiches on pizza or for salad bars

The results of the present study revealed that thenonthermal process (modified method 6) for pepperonimanufacture was not sufficient to effect a 5-10g unit reduc-tion in counts of E coli 0157H7 The observed lt2-10g unitdecrease in counts of the pathogen following fermentation ofpepperoni at 96degF (36degC) to pH ~ 50 and drying was inagreement with a previous report of a 1- to 2-10g unitdecrease in counts of the pathogen under similar processconditions for salami (ie fermentation at 96degF (36degC) topH 48 drying to a MPr of ~191 and storage at 4degC for 2weeks) (15) In the present study postfermentation heatingachieved a 5-D kill of this pathogen Although both heatingregimens evaluated were effective for eliminating E coliOI57H7 cooking at l28degF (53degC) for 60 min was lessharmful to the product than cooking to an internal tempera-ture of 145degF (63degC) These data represent the first of severalreports expected to be published on the fate of E coli0157H7 in fermented meats by our group and by otherinvestigators In addition to a thermal processing step weare evaluating the eff~ct of fermentation temperature andpH a 3- to 7-day postfermentation storage at low pH andvarious heating regimens for eliminating this pathogen inready-to-eat fermented sausage It is possible that slightincreases in the fermentation temperature andor extendedstorage at low pH would result in a greater reduction incounts of the pathogen and further reduce the time andtemperature required for postfermentation heating A compre-hensive effort to screen raw materials for the pathogen tofollow good manufacturing practices and to monitor fin-ished product would significantly reduce the likelihood of

illness due to E coli 0157H7 in fermented sausage In theinterim the results of the present study will assist manufac-turers in prioritizing options for validating the safety of theirrespective manufacturing processes

ACKNOWLEDGMENTS

The technical assistance of Patrick Krakar Tamara Dodge and KarenArnold of the Food Research Institute and Rong Chen and Jane Wagner ofDoskocil Foods are gratefully acknowledged Our appreciation is alsoextended to Chuck Baum (OW-Madison Biotron) John Cerveny (OscarMayer) Steven Devcich (TeePak) Gerry Durnell (National Association ofPizza Operators) Larry Hand (Diversitech) Steve Lary (Alkar) and SteveSeideman (Doskocil) for supplies service equipment andor consultationsWe also thank the following individuals for supplying cultures Jay HLewis (Washington State Public Health Lab) for strain F-90 Mary EProctor (Wisconsin Bureau of Public Health) for strain SLH21788 KathyGlass (Food Research Institute) for strain C7927 and Joy G Wells andTimothy J Barrett (Centers for Disease Control and Prevention) for strainC9490

This work was supported by a gift from Doskocil Companies Inc andthe College of Agricultural and Life Sciences of the University ofWisconsin-Madison

REFERENCES

1 Abdul-Raouf U M L R Beuchat and M S Ammar 1993 Survivaland growth of Escherichia coli 0157H7 on salad vegetables ApplEnviron Microbiol 59 1999-2006

2 Abdul-Raouf U M L R Beuchat and M S Ammar 1993 Survivaland growth of Escherichia coli 0157H7 in ground roasted beef asaffected by pH acidulants and temperature Appl Environ Micro-bioI 592364-2368

3 Arnold K W and C W Kaspar 1995 Starvation- and stationary-phase-induced acid tolerance in Escherichia coli 0157H7 ApplEnviron Microbiol 612037-2039

4 Arocha M M M McVey S D Loder J H Rupnow and L BBullerman 1992 Behavior of hemorrhagic Escherichia coli 0157H7during the manufacture of cottage cheese J Food Prot 55379-381

5 Benjamin M M andA R Datta 1995 Acid tolerance of enterohem-orrhagic Escherichia coli Appl Environ Microbiol 61 1669-1672

6 Besser R E S M Lett J T Weber M P Doyle T J Barrett J GWells and P M Griffin 1993 An outbreak of diarrhea and hemolyticuremic syndrome from Escherichia coli 0157H7 in fresh-pressedapple cider JAMA 2692217-2220

7 Centers for Disease Control and Prevention 1993 Update multi stateoutbreak of Escherichia coli 0157H7 infections from hamburgers-western United States 1992-1993 Morbid Mortal Weekly Rep42258-263

8 Centers for Disease Control and Prevention 1995 Escherichia coli0157H7 outbreak linked to commercially distributed dry-curedsalami-Washington and California 1994 Morbid Mortal WeeklyRep 44157-160

9 Centers for Disease Control and Prevention 1995 Communityoutbreak of hemolytic uremic syndrome attributable to Escherichiacoli 0111 NM-South Australia 1995 Morbid Mortal Weekly Rep44550-551 557

10 Conner D E and J S~ Kotrola 1995 Growth and survival ofEscherichia coli 0157H7 under acidic conditions Appl EnvironMicrobiol 61382-385

l1 Cutter C N and G R Siragusa 1994 Efficacy of organic acidsagainst Escherichia coli 0157H7 attached to beef carcass tissue usinga pilot scale model carcass washer J Food Prot 5797-103

12 Darmadjii P M Izumimoto T Miyamoto and K Katoaka 1990Lactic fermentation on preservative qualities of Dendeng giling JFood Sci 55(6) 1523-1527

13 Doyle M P and J L Schoeni 1984 Survival and growthcharacteristics of Escherichia coli associated with hemorrhagiccolitis Appl Environ Microbiol 48855-856

14 Faith N G J Hinkens T D Lorang P Bailey D Buege C WKaspar and J B Luchansky 1996 Validation of pepperoni processesfor control of Escherichia coli 0157H7 abstr 191 p 80 Progr

1264 HINKENS FAITH LORANG BAILEY BUEGE KASPAR AND LUCHANSKY

temperature was then raised at a rate of 12 to 15degF (70 toSoC) per h until an internal chub temperature of 145degF(63degC) was reached (Table 2) Heating to 145degF (63degC)reduced counts of the pathogen below detection (gt50-logunit decrease) however it was possible to recover viablecells of serotype 0157 H7 by enrichment after heating andup to S days of drying Nevertheless heating chubs to aninternal temperature of 145degF (63degC) delivered the required5-log unit decrease in counts of the pathogen withoutcausing any visible textural or color changes to the finalproduct (data not shown)

Fermentation at 96degF (36degC) to pHS 50 heating to aninternal temperature of 128degF (53degC) for 60 min anddrying at 55degF (13degC) and 65 RH to a MIProfS161

On the basis of the encouraging results with the 145degF(63degC) heating process a second heating regimen wasevaluated As in the previous experiments fermentationalone had no appreciable effect on the pathogen (09-logcfug decrease) (Table 3) However holding the chubs at aninternal temperature of 12soF (53degC) for 60 min (Table 3)resulted in at least a 5-log CFUg reduction in numbers of0157H7 After heating it was only possible to recover thepathogen by enrichment Also heating for 60 min at 12soF(53degC) did not visibly affect the appearance of the resultingproduct and did not cause greasing out (ie there was noresidual fat layer under the casing) (data not shown)

DISCUSSION

Escherichia coli 0157H7 emerged as a foodbornepathogen of primary health concern in the 1990s During thisperiod serotype 0157H7 strains were incriminated innumerous sporadic cases and several outbreaks involving avariety of foods (6 7-9 22 29) Foods of bovine origin ingeneral and ground beef in particular were most oftenidentified as the food vehicles (16) To better understand theresponse of this bacterium to intrinsic (eg pH) andextrinsic (eg temperature) conditions associated with foodsstudies were conducted using apple cider (6 20) cottagecheese (4) ground beef (13 17 26) mayonnaise andordressings (23 29) and salad vegetables (1) In general thesestudies revealed E coli 0157 H7 is very tolerant to acidAlthough there is significant variation in the degree of acidtolerance from strain to strain most 0157H7 strains aremore acid tolerant than other E coli (3 5 11 20) HoweverE coli 0157H7 does not appear to be particularly heattolerant (13) Although there have been reports on thebehavior of non-0157H7 strains of E coli in fermentedmeats (12 28) with the exception of a previous studyconducted at the Food Research Institute using salami (15)no information has been published on the behavior ofserotype 0157H7 strains in fermented sausages

Several factors associated with pepperoni batter wereevaluated for possible effects on the viability of the serotype0157H7 cocktail For example high levels of indigenousbacteria could have a negative impact on the behavior of Ecoli 0157H7 However the relatively low numbers ofaerobic bacteria in the raw materials suggest that indigenous

bacteria were probably not a significant factor in this studyLikewise microbiological analyses of the batter revealedthat the spice and cure mixtures had no significant effect onthe 0157H7 cocktail Due to the limitations of the facilityhousing the smokehouse it was not possible to evaluate theeffect of smoke in this study Smoking is presumably mosteffective against surface contamination of sausage and assuch was not expected to have an appreciable effect on cellsof E coli within pepperoni chubs The addition of smokecould enhance the kill achieved by fermentation but itwould probably not be sufficient to deliver a 5-D killAmong the variety of other factors that could influence theviability of serotype 0157H7 strains in pepperoni the pHand acid type as well as the time held at a given temperatureprobably present the greatest potential for impacting patho-gen numbers (2) For example E coli 0157H7 is moresensitive to acid conditions (pH 36 to 40) at abuse (15 to20degC or 60 to 77degF) rather than refrigeration (4 to SoC or 39to 46degF) temperatures (6 20) It is also significant thatsurvival of serotype 0157H7 strains can also be affected bythe type of acid In a study by Abdul-Raouf et al (2)evaluating the effect of acidulants on E coli 0157H7 inbeef slurries the order of effectiveness for inhibiting growthwas acetic gt lactic gt citric acid at all pH values andtemperatures tested However Conner and Kotrola (10)reported that organic acids and salts enhanced the survival ofserotype 0157H7 strains during storage at 4degC in asynthetic medium These data underscore the importance ofconducting validation studies in bona fide fermented meatsrather than microbiological media andor beaker sausage tomore precisely establish the viability of E coli 0157H7 fora given formulation and manufacturing process

Although lower pH andor higher fermentation tempera-tures may result in a greater decrease in pathogen numbersit is unlikely that these parameters could be adjustedsufficiently to deliver a 5-D kill in pepperoni More specifi-cally at pH S 43 pepperoni may be too tart for mostconsumers Similarly due to intrinsic limitations of thestarter culture it may not be possible to conduct a fermenta-tion at temperatures much in excess of 120degF (49degC) Forthese reasons it was necessary to evaluate postfermentationheating of chubs as a method to deliver the mandated 5-Dreduction in counts of E coli 0157H7 The first thermalprocess tested 145degF (63degC) internal was selected becauseit meets the requirements for trichinae destruction andbecause it approximates the time temperature and humidityguidelines established by the USDA for cooked andor roastbeef (27) The cooking requirements for cooked andor roastbeef were also approved by the USDA FSIS as an option tocontrol serotype 0157H7 strains of E coli in fermentedsausage (25) The other thermal process tested 12soF(53degC) internal for 60 min was selected because it alsomeets the requirements for trichinae control but moreimportant it is less harmful to the texture and appearance ofpepperoni than the 145degF (63degC) cooking regimen (SteveSeideman personal communication) This approach is alsonot without limitations as some sausage varieties are notamenable to postfermentation heating due to possible unde-

E COLI 0157H7 IN PEPPERONI 1265

sirable textural effects on the product andor some manufac-turers are not capable of including such a heating step due tothe volume of product produced or the cost of the attendantequipment and facilities From a practical standpoint heat-ing may cause greasing out and may yield a product moreprone to cupping when used as a pizza topping (SteveSeideman personal communication) Further studies may bewarranted to identify heating regimens that are sufficient toeliminate the pathogen without producing untoward effectson the finished product and that are cost effective

The present study evaluated the modified method 6pepperoni process as well as postfermentation heating toinactivate E coli 0157H7 A comprehensive sampling andscreening of raw materials to account for low levels of thispathogen in raw materials as well as adherence to goodmanufacturing procedures would also contribute to thesafety of fermented meat products As another considerationthe sampling plan for raw materials that will undergosubsequent heat processing such as pepperoni used as pizzatopping is less stringent (ie 15 samples per lot comparedto 30 samples per lot [25 J) because the ensuing cooking ispresumably sufficient to eliminate the pathogen Perhaps amore serious threat to consumer safety would result fromcontaminated pepperoni used for sandwiches or salad barsFor these uses it is necessary to include a postfermentationheating step sufficient to deliver the 5-D kill Thus it wouldalso be informative to substantiate the fate of the pathogenafter frozen or refrigerated storage of pepperoni targeted foruse in sandwiches on pizza or for salad bars

The results of the present study revealed that thenonthermal process (modified method 6) for pepperonimanufacture was not sufficient to effect a 5-10g unit reduc-tion in counts of E coli 0157H7 The observed lt2-10g unitdecrease in counts of the pathogen following fermentation ofpepperoni at 96degF (36degC) to pH ~ 50 and drying was inagreement with a previous report of a 1- to 2-10g unitdecrease in counts of the pathogen under similar processconditions for salami (ie fermentation at 96degF (36degC) topH 48 drying to a MPr of ~191 and storage at 4degC for 2weeks) (15) In the present study postfermentation heatingachieved a 5-D kill of this pathogen Although both heatingregimens evaluated were effective for eliminating E coliOI57H7 cooking at l28degF (53degC) for 60 min was lessharmful to the product than cooking to an internal tempera-ture of 145degF (63degC) These data represent the first of severalreports expected to be published on the fate of E coli0157H7 in fermented meats by our group and by otherinvestigators In addition to a thermal processing step weare evaluating the eff~ct of fermentation temperature andpH a 3- to 7-day postfermentation storage at low pH andvarious heating regimens for eliminating this pathogen inready-to-eat fermented sausage It is possible that slightincreases in the fermentation temperature andor extendedstorage at low pH would result in a greater reduction incounts of the pathogen and further reduce the time andtemperature required for postfermentation heating A compre-hensive effort to screen raw materials for the pathogen tofollow good manufacturing practices and to monitor fin-ished product would significantly reduce the likelihood of

illness due to E coli 0157H7 in fermented sausage In theinterim the results of the present study will assist manufac-turers in prioritizing options for validating the safety of theirrespective manufacturing processes

ACKNOWLEDGMENTS

The technical assistance of Patrick Krakar Tamara Dodge and KarenArnold of the Food Research Institute and Rong Chen and Jane Wagner ofDoskocil Foods are gratefully acknowledged Our appreciation is alsoextended to Chuck Baum (OW-Madison Biotron) John Cerveny (OscarMayer) Steven Devcich (TeePak) Gerry Durnell (National Association ofPizza Operators) Larry Hand (Diversitech) Steve Lary (Alkar) and SteveSeideman (Doskocil) for supplies service equipment andor consultationsWe also thank the following individuals for supplying cultures Jay HLewis (Washington State Public Health Lab) for strain F-90 Mary EProctor (Wisconsin Bureau of Public Health) for strain SLH21788 KathyGlass (Food Research Institute) for strain C7927 and Joy G Wells andTimothy J Barrett (Centers for Disease Control and Prevention) for strainC9490

This work was supported by a gift from Doskocil Companies Inc andthe College of Agricultural and Life Sciences of the University ofWisconsin-Madison

REFERENCES

1 Abdul-Raouf U M L R Beuchat and M S Ammar 1993 Survivaland growth of Escherichia coli 0157H7 on salad vegetables ApplEnviron Microbiol 59 1999-2006

2 Abdul-Raouf U M L R Beuchat and M S Ammar 1993 Survivaland growth of Escherichia coli 0157H7 in ground roasted beef asaffected by pH acidulants and temperature Appl Environ Micro-bioI 592364-2368

3 Arnold K W and C W Kaspar 1995 Starvation- and stationary-phase-induced acid tolerance in Escherichia coli 0157H7 ApplEnviron Microbiol 612037-2039

4 Arocha M M M McVey S D Loder J H Rupnow and L BBullerman 1992 Behavior of hemorrhagic Escherichia coli 0157H7during the manufacture of cottage cheese J Food Prot 55379-381

5 Benjamin M M andA R Datta 1995 Acid tolerance of enterohem-orrhagic Escherichia coli Appl Environ Microbiol 61 1669-1672

6 Besser R E S M Lett J T Weber M P Doyle T J Barrett J GWells and P M Griffin 1993 An outbreak of diarrhea and hemolyticuremic syndrome from Escherichia coli 0157H7 in fresh-pressedapple cider JAMA 2692217-2220

7 Centers for Disease Control and Prevention 1993 Update multi stateoutbreak of Escherichia coli 0157H7 infections from hamburgers-western United States 1992-1993 Morbid Mortal Weekly Rep42258-263

8 Centers for Disease Control and Prevention 1995 Escherichia coli0157H7 outbreak linked to commercially distributed dry-curedsalami-Washington and California 1994 Morbid Mortal WeeklyRep 44157-160

9 Centers for Disease Control and Prevention 1995 Communityoutbreak of hemolytic uremic syndrome attributable to Escherichiacoli 0111 NM-South Australia 1995 Morbid Mortal Weekly Rep44550-551 557

10 Conner D E and J S~ Kotrola 1995 Growth and survival ofEscherichia coli 0157H7 under acidic conditions Appl EnvironMicrobiol 61382-385

l1 Cutter C N and G R Siragusa 1994 Efficacy of organic acidsagainst Escherichia coli 0157H7 attached to beef carcass tissue usinga pilot scale model carcass washer J Food Prot 5797-103

12 Darmadjii P M Izumimoto T Miyamoto and K Katoaka 1990Lactic fermentation on preservative qualities of Dendeng giling JFood Sci 55(6) 1523-1527

13 Doyle M P and J L Schoeni 1984 Survival and growthcharacteristics of Escherichia coli associated with hemorrhagiccolitis Appl Environ Microbiol 48855-856

14 Faith N G J Hinkens T D Lorang P Bailey D Buege C WKaspar and J B Luchansky 1996 Validation of pepperoni processesfor control of Escherichia coli 0157H7 abstr 191 p 80 Progr

E COLI 0157H7 IN PEPPERONI 1265

sirable textural effects on the product andor some manufac-turers are not capable of including such a heating step due tothe volume of product produced or the cost of the attendantequipment and facilities From a practical standpoint heat-ing may cause greasing out and may yield a product moreprone to cupping when used as a pizza topping (SteveSeideman personal communication) Further studies may bewarranted to identify heating regimens that are sufficient toeliminate the pathogen without producing untoward effectson the finished product and that are cost effective

The present study evaluated the modified method 6pepperoni process as well as postfermentation heating toinactivate E coli 0157H7 A comprehensive sampling andscreening of raw materials to account for low levels of thispathogen in raw materials as well as adherence to goodmanufacturing procedures would also contribute to thesafety of fermented meat products As another considerationthe sampling plan for raw materials that will undergosubsequent heat processing such as pepperoni used as pizzatopping is less stringent (ie 15 samples per lot comparedto 30 samples per lot [25 J) because the ensuing cooking ispresumably sufficient to eliminate the pathogen Perhaps amore serious threat to consumer safety would result fromcontaminated pepperoni used for sandwiches or salad barsFor these uses it is necessary to include a postfermentationheating step sufficient to deliver the 5-D kill Thus it wouldalso be informative to substantiate the fate of the pathogenafter frozen or refrigerated storage of pepperoni targeted foruse in sandwiches on pizza or for salad bars

The results of the present study revealed that thenonthermal process (modified method 6) for pepperonimanufacture was not sufficient to effect a 5-10g unit reduc-tion in counts of E coli 0157H7 The observed lt2-10g unitdecrease in counts of the pathogen following fermentation ofpepperoni at 96degF (36degC) to pH ~ 50 and drying was inagreement with a previous report of a 1- to 2-10g unitdecrease in counts of the pathogen under similar processconditions for salami (ie fermentation at 96degF (36degC) topH 48 drying to a MPr of ~191 and storage at 4degC for 2weeks) (15) In the present study postfermentation heatingachieved a 5-D kill of this pathogen Although both heatingregimens evaluated were effective for eliminating E coliOI57H7 cooking at l28degF (53degC) for 60 min was lessharmful to the product than cooking to an internal tempera-ture of 145degF (63degC) These data represent the first of severalreports expected to be published on the fate of E coli0157H7 in fermented meats by our group and by otherinvestigators In addition to a thermal processing step weare evaluating the eff~ct of fermentation temperature andpH a 3- to 7-day postfermentation storage at low pH andvarious heating regimens for eliminating this pathogen inready-to-eat fermented sausage It is possible that slightincreases in the fermentation temperature andor extendedstorage at low pH would result in a greater reduction incounts of the pathogen and further reduce the time andtemperature required for postfermentation heating A compre-hensive effort to screen raw materials for the pathogen tofollow good manufacturing practices and to monitor fin-ished product would significantly reduce the likelihood of

illness due to E coli 0157H7 in fermented sausage In theinterim the results of the present study will assist manufac-turers in prioritizing options for validating the safety of theirrespective manufacturing processes

ACKNOWLEDGMENTS

The technical assistance of Patrick Krakar Tamara Dodge and KarenArnold of the Food Research Institute and Rong Chen and Jane Wagner ofDoskocil Foods are gratefully acknowledged Our appreciation is alsoextended to Chuck Baum (OW-Madison Biotron) John Cerveny (OscarMayer) Steven Devcich (TeePak) Gerry Durnell (National Association ofPizza Operators) Larry Hand (Diversitech) Steve Lary (Alkar) and SteveSeideman (Doskocil) for supplies service equipment andor consultationsWe also thank the following individuals for supplying cultures Jay HLewis (Washington State Public Health Lab) for strain F-90 Mary EProctor (Wisconsin Bureau of Public Health) for strain SLH21788 KathyGlass (Food Research Institute) for strain C7927 and Joy G Wells andTimothy J Barrett (Centers for Disease Control and Prevention) for strainC9490

This work was supported by a gift from Doskocil Companies Inc andthe College of Agricultural and Life Sciences of the University ofWisconsin-Madison

REFERENCES

1 Abdul-Raouf U M L R Beuchat and M S Ammar 1993 Survivaland growth of Escherichia coli 0157H7 on salad vegetables ApplEnviron Microbiol 59 1999-2006

2 Abdul-Raouf U M L R Beuchat and M S Ammar 1993 Survivaland growth of Escherichia coli 0157H7 in ground roasted beef asaffected by pH acidulants and temperature Appl Environ Micro-bioI 592364-2368

3 Arnold K W and C W Kaspar 1995 Starvation- and stationary-phase-induced acid tolerance in Escherichia coli 0157H7 ApplEnviron Microbiol 612037-2039

4 Arocha M M M McVey S D Loder J H Rupnow and L BBullerman 1992 Behavior of hemorrhagic Escherichia coli 0157H7during the manufacture of cottage cheese J Food Prot 55379-381

5 Benjamin M M andA R Datta 1995 Acid tolerance of enterohem-orrhagic Escherichia coli Appl Environ Microbiol 61 1669-1672

6 Besser R E S M Lett J T Weber M P Doyle T J Barrett J GWells and P M Griffin 1993 An outbreak of diarrhea and hemolyticuremic syndrome from Escherichia coli 0157H7 in fresh-pressedapple cider JAMA 2692217-2220

7 Centers for Disease Control and Prevention 1993 Update multi stateoutbreak of Escherichia coli 0157H7 infections from hamburgers-western United States 1992-1993 Morbid Mortal Weekly Rep42258-263

8 Centers for Disease Control and Prevention 1995 Escherichia coli0157H7 outbreak linked to commercially distributed dry-curedsalami-Washington and California 1994 Morbid Mortal WeeklyRep 44157-160

9 Centers for Disease Control and Prevention 1995 Communityoutbreak of hemolytic uremic syndrome attributable to Escherichiacoli 0111 NM-South Australia 1995 Morbid Mortal Weekly Rep44550-551 557

10 Conner D E and J S~ Kotrola 1995 Growth and survival ofEscherichia coli 0157H7 under acidic conditions Appl EnvironMicrobiol 61382-385

l1 Cutter C N and G R Siragusa 1994 Efficacy of organic acidsagainst Escherichia coli 0157H7 attached to beef carcass tissue usinga pilot scale model carcass washer J Food Prot 5797-103

12 Darmadjii P M Izumimoto T Miyamoto and K Katoaka 1990Lactic fermentation on preservative qualities of Dendeng giling JFood Sci 55(6) 1523-1527

13 Doyle M P and J L Schoeni 1984 Survival and growthcharacteristics of Escherichia coli associated with hemorrhagiccolitis Appl Environ Microbiol 48855-856

14 Faith N G J Hinkens T D Lorang P Bailey D Buege C WKaspar and J B Luchansky 1996 Validation of pepperoni processesfor control of Escherichia coli 0157H7 abstr 191 p 80 Progr

1266 HINKENS FAITH LORANG BAILEY BUEGE KASPAR AND LUCHANSKY

Abstr Book 83rd Annu Meet IAMFES International Association ofMilk Food and Environmental Sanitarians Des Moines IA

15 Glass K A J M Loeffelholz J P Ford and M P Doyle 1992 Fateof Escherichia coli 0157H7 as affected by pH or sodium chloride andin fermented dry sausage App Environ Microbio 582513-2516

16 Griffin P M 1995 Escherichia coli 0157H7 and other enterohemor-rhagic Escherichia coli p 739-761 In M J Blaser P D Smith J 1Ravdin H B Greenberg and R L Guerrant (ed) Infections of thegastrointestinal tract Raven Press New York

17 Line J E A R Fain Jr A B Moran L M Martin R VLechowich J M Carosella and W L Brown 1991 Lethality of heatto Escherichia coli 0157H7 D-value and Z-value determinations inground beef J Food Prot 54762-766

18 Luchansky J B K A Glass K D Harsono A J Degnan N GFaith B Cauvin G Baccus-Taylor K Arihara B Bater A JMaurer and R G Cassens 1992 Genomic analysis of Pediococcusstarter cultures used to control Listeria monocytogenes in turkeysummer sausage App Environ Microbio 583053-3059

19 McNeal J E 1990 Meat and meat products p 931-938 In KHerlich (ed) Official methods of analysis 15th ed Association ofOfficial Analytical Chemists Arlington VA

20 Miller L G and C W Kaspar 1994 Escherichia coli 0157H7 acidtolerance and survival in apple cider J Food Prot 57460--464

21 Okrend A J G B E Rose and R Matner 1990 An improvedscreening method for the detection and isolation of Escherichia coli0157H7 from meat incorporating the 3M Petrifilml test kit-EHECfor hemorrhagic Escherichia coli 0157H7 J Food Prot 53936-940

22 Proctor M E 1996 Surveillance for Escherichia coli 0157H7

infection in Wisconsin 1993-1995 investigation of three outbreaksWisconsin Epidemio Bull 17(1) 1-4

23 Raghubeer E V J S Ke M L Campbell and R S Meyer 1995Fate of Escherichia coli 0157H7 and other coliforms in commercialmayonnaise and refrigerated salad dressing J Food Prot 58 13-18

24 Reed C A 1995 Challenge study-Escherichia coli 0157H7 infermented sausage US Department of Agriculture Food Safety andInspection Service Washington DC April 28 1995 letter to PlantManagers

25 Reed C A 1995 Approaches for ensuring the safety of dry andsemi-dry fermented sausage products US Department of Agricul-ture Food Safety and Inspection Service Washington DC August21 1995 letter to Plant Managers

26 Todd E A Hughes J MacKenzie R Caldeira T Gleeson and BBrown 1993 Thermal resistance ofverotoxigenic Escherichia coli inground beef-initial work p 93-109 In E C D Todd and J MMackenzie (ed) Escherichia coli and other verocytotoxigenic E coliin foods Health and Welfare Canada Ottawa Canada

27 United States Department of Agriculture Food Safety and InspectionService 1994 Code of Federal Regulations Title 9 31810 Office ofthe Federal Register US Government Printing Office WashingtonDC

28 Unluturk A and F Turantas 1991 Fate of coliforms in TurkishSoudjuk during ripening and storage J Sci Food Agric 57399-404

29 Weagent S D J L Bryant and D H Bark 1994 Survival ofEscherichia coli 0157H7 in mayonnaise and mayonnaise-basedsauces at room and refrigerated temperatures J Food Prot 57 629-631