Surveillance and outbreak reports

Transcription

1 Surveillnce nd outbrek reports Outbrek of slmonellosis fter wedding prty, Bvri, Germny, summer 2010: the importnce of implementing food sfety concepts B von Wissmnn (betrix.wissmnn@gmil.com) 1,2, C Klinc 1, R Schulze 3, A Wolf 3, H Schreiner 4, W Rbsch 5, R Prger 5, W Hutmnn 1 1. Bvrin Helth nd Food Sfety Authority, Helth, Oberschleissheim, Germny 2. Postgrdute Fellowship in Applied Epidemiology, Robert Koch Institute, Berlin, Germny 3. Locl helth office Ansbch, Ansbch, Germny 4. Bvrin Helth nd Food Sfety Authority, Food Sfety nd Cosmetics, Erlngen, Germny 5. Ntionl Reference Centre for Slmonell nd other Enteric pthogens, Robert Koch Institute, Wernigerode Brnch, Bcteril Infections, Wernigerode, Germny Cittion style for this rticle: von Wissmnn B, Klinc C, Schulze R, Wolf A, Schreiner H, Rbsch W, Prger R, Hutmnn W. Outbrek of slmonellosis fter wedding prty, Bvri, Germny, summer 2010: the importnce of implementing food sfety concepts. Euro Surveill. 2012;17(6):pii= Avilble online: Article published on 9 Februry 2012 An outbrek of slmonellosis in the summer of 2010 fter wedding prty in Bvri, Germny, ws investigted, to identify vehicles nd source of the outbrek nd ny flws in food sfety procedures. A cohort study trgeting ll wedding prticipnts ws crried out in ddition to lbortory nd environmentl investigtions. The overll ttck rte ws 59% (52/88). A group of buffet dishes contining myonnise ws ssocited with the highest reltive risk: 6.6 (95% CI: ). Subtyping of subset of Slmonell isoltes cultured from stool smples from wedding prticipnts nd ctering stff nd from food smples showed the isoltes to be indistinguishble (S. Enteritidis, lysotype 4/6, ribotype 3). Europen lws require food business opertors to estblish nd implement food sfety procedures. Investigtion of the prctices of the ctering compny reveled lck of stff trining nd the bsence of records of food sfety concept. Nonedible flower decortions in cold dishes were identified s potentil source of secondry contmintion becuse their cultivtion nd processing re not subject to the food sfety regultions minimising the risk of contmintion of crops intended for humn consumption. The outbrek demonstrted the importnce of stff trining nd food sfety concepts to identify nd mitigte hzrds to food sfety. Introduction Slmonellosis is n importnt cuse of food-borne gstroenteritis in Europe [1]. Prophylctic mesures re imed t ll stges of food production from frm to fork nd re set down in Europen lws governing food sfety. Complince of food business opertors to the minimum hygiene regultions is checked through officil controls. Additionlly, food business opertors re required to estblish nd operte food sfety procedures bsed on the hzrd nlysis nd criticl control points (HACCP) principles [2]. These principles form flexible but structured pproch to identifying hzrds (e.g. risk of contmintion) s well s control points (e.g. heting or cooling steps in the production or preprtion of foods) tht cn mitigte theses hzrds. The procedures developed hve to be documented by the food business opertor nd regulr checks of the control mesures hve to be recorded (e.g. temperture checks to record the mintennce of the cold chin). While they do not replce officil controls, the HACCP principles re centrl to the Europen concept of food sfety by helping food business opertors to ttin high stndrd of food sfety. Investigtions of foodborne outbreks serve to identify the source of the outbreks nd my help to identify flws or wek points in the plnning nd implementtion of food hygiene mesures. The results my lso be used to ugment generl guidelines for good prctice in hygiene nd food sfety. Surveillnce of slmonellosis in Germny is bsed on mndtory notifiction of lbortory-confirmed infections by the lbortory to the locl helth office [3]. Additionlly, the locl helth office my be lerted to cluster of clinicl cses with suspected common source by clinicin or n ffected member of the public. The locl helth office further investigtes the reported cses, crries out ctive cse finding s prt of the outbrek investigtion nd implements control mesures. Cse reports re forwrded to the federl nd ntionl level using n electronic dtbse for notifible infectious diseses. Isolted Slmonell strins cn be sent to the Germn Ntionl Reference Centre for Slmonell nd other Enteric pthogens t 1

2 the Robert Koch Institute for subtyping by clssicl nd moleculr methods. In the summer of 2010, locl helth office in northern Bvri, Germny, ws informed tht pproximtely hlf of the 110 guests of wedding tht hd tken plce the preceding weekend hd contrcted gstroenteritis. At the wedding prty, soup nd lte-lunch buffet (served from 3 p.m.) nd cold dinner buffet (served from 10 p.m.) hd been provided by n out-oftown cterer. In ddition, wedding cke mde by locl bkery nd number of ckes nd desserts contributed by different wedding guests were served by the ctering stff t The food served t the wedding ws suspected to be the source of the outbrek. Initil lbortory results of stool smples of some guests who becme ill indicted Slmonell enteric subsp. enteric serovr Enteritidis (S. Enteritidis) s the custive pthogen. The locl helth office, supported by the Bvrin Helth nd Food Sfety Authority, initited comprehensive outbrek investigtion including cohort study, pthogen subtyping nd n environmentl investigtion ws conducted to identify the vehicle nd source of the outbrek. Methods Cohort study The study popultion for the cohort study ws defined s ll people who ttended the wedding prty nd te ny of the served food. Completion of written questionnire ws prerequisite for individuls to be included in the cohort study. Ctering personnel were excluded from the cohort study due to perceived conflict of interest (the compny ws fcing litigtion). Written questionnire A structured, written questionnire nd stmped ddressed envelope for its return were posted to ll people who ttended the wedding prty nd for whom complete postl ddress could be obtined, two weeks fter the wedding. The questionnire included stndrdised questions relting to demogrphy nd the development of symptoms of slmonellosis, including dte nd time of onset, s used in previous outbrek investigtions conducted by the Bvrin Helth nd Food Sfety Authority. Questions on food consumption were dpted to refer to the specific food items served t the wedding prty. Cse definition A clinicl cse of slmonellosis ws person who hd eten t the wedding prty nd who developed t lest one symptom of slmonellosis (nuse, vomiting, stomch crmps, dirrhoe (three or more loose stools in 24 hours), pyrexi (body temperture bove 38.5 C)) between six hours fter the strt of the buffet (which begn t 3 p.m.) nd 72 hours fter the end of the buffet (the end ws considered to be t m., when the cterers left). Lbortory confirmtion of slmonell infection ws obtined by the isoltion of Slmonell species from stool smple tken s prt of the outbrek investigtion, irrespective of the presence or bsence of symptoms of slmonellosis. People with clinicl nd/or lbortory-confirmed slmonell infections were considered s slmonellosis cses in the cohort study. Sttisticl nlysis For the descriptive nlysis of cses nd non-cses in the cohort study, continuous vribles were summrised using medin nd rnge. Ctegoricl vribles were presented using bsolute numbers nd percentges. Where pproprite, binomil 95% confidence intervl (CI) ws clculted for the percentges. The univrible ssocition between cse sttus nd exposure to individul dishes ws tested using the chisqure test or Fisher s exct test where pproprite. The effect of ech exposure vrible ws quntified by estimtion of the reltive risk nd the respective 95% CI. When the numbers were smll, exct Poisson regression ws used for estimtion. Multivrible nlysis ws performed using generlised liner model with log link nd binomil error distribution. Exposures with p vlue of <0.1 were included in the multivrible model. The multivrible model ws constructed by first including ll vribles with p vlue of <0.1 nd then dropping vribles mnully in bckwrds elimintion procedure bsed on the Wld test. Sttisticl significnce ws ssumed when p<0.05 nd the 95% CI of the reltive risk did not include 1. When interprettion of the p vlue nd the 95% CI were contrdictory, the more conservtive interprettion (not sttisticlly significnt) ws reported. Dt nlysis ws crried out in STATA Grouping for nlysis purposes Dishes served t the wedding prty were grouped by the four different courses served: soup, lunch buffet, ckes nd dinner buffet. Further groupings of the food items were creted for nlysis ccording to common ingredients (e.g. one group for ll met dishes tht were served in crem suce, nother group for ll dishes contining myonnise) or ccording to the type of food or preprtion nd the perceived risk of contmintion (e.g. ll tinned food items were put into one group, s their probbility of contmintion with Slmonell ws similrly low). A Cochrn-Armitge test for trend ws performed for the ssocition between the risk of infection nd the number of dishes contining myonnise consumed, to ssess the dose-response. Environmentl investigtion Although the ctering stff were excluded from the cohort study, ll eight stff members (one of whom ws not present t the wedding prty but ws involved in preprtion of the dishes) were contcted by telephone by the locl helth office on the dy the office ws notified bout the outbrek (four dys fter the wedding event), to ensure tht individuls with symptoms of slmonellosis did not continue to hndle food nd to determine whether ny of the ctering stff hd displyed symptoms of gstroenteritis before the 2

3 wedding event. The ctering compny s records for stff hygiene trining s well s their records for food temperture controls (s n indictor vrible for the implementtion of HACCP procedures) were checked. The kitchen fcilities of the venue of the wedding prty were inspected nd swbs of the work surfces were sent for lbortory testing. The premises of the ctering compny nd their fcilities were inspected by the responsible locl veterinry office nd swbs of the work surfces were submitted for lbortory testing. All wedding guests were contcted by the locl helth office by telephone four dys fter the wedding to identify clinicl cses of slmonellosis nd to exclude such person with slmonellosis working in the food industry from hndling food until they were proven to be no longer infected. Food leftovers tken home by wedding guests were submitted for lbortory dignosis, s no food items were left t the prty venue or t the cterer s fcilities. Lbortory dignosis Stool smples Stool smples from ll ctering personnel (n=8) were submitted to the locl helth office nd tested for Slmonell nd Shigell (by culture) nd for norovirus (by polymerse chin rection (PCR)) in medicl lbortory. Positive smples were forwrded to the Ntionl Reference Centre for Slmonell nd other Enteric pthogens t the Robert Koch Institute. Stool smples from wedding guests were submitted to medicl lbortories s prt of the clinicl dignosis by their generl prctitioners (GPs). The smples were tested for Slmonell nd Shigell (by culture) nd for other pthogens (Cmpylobcter nd Yersini enterocolitic (by culture) nd norovirus (by PCR)), depending on the request submitted by the GP. For number of smples, Slmonell isoltes were serotyped (depending on the lbortory). At the request of the locl helth office, Slmonell-positive smples or isoltes were forwrded to the Ntionl Reference Centre if the mteril ws still vilble. Food smples nd swbs Food smples nd swbs of the kitchen fcilities t the wedding prty venue, s well s swbs of the cterer s kitchen fcilities were tested for Slmonell (by culture) nd norovirus (by PCR). Slmonell isoltes were serotyped nd forwrded to the Ntionl Reference Centre. Subtyping of isoltes At the Ntionl Reference Centre, subsmple of the Slmonell isoltes from stool smples nd food smples were phge typed nd ribotyped. The routine phge typing of S. Enteritidis ws performed ccording to the typing system of Wrd et l. [4] s well s tht of Llko nd László [5], e.g. phge type (PT) 4/6, i.e. phge type 4 ccording to Wrd et l. nd 6 ccording to Llko nd Lszlo. The ribotyping ws performed ccording to Liebn et l. [6] nd Rbsch et l. [7]. Results Response to questionnire Questionnires were sent to ll wedding guests (including the bride nd groom) for whom complete postl ddress ws obtined (n=100). Of these, 91 returned their completed questionnires. Among the 19 people for whom questionnire dt were not vilble (10 people for whom no vlid ddress ws vilble, nine people who did not return the questionnire), five stted during their telephone interview with the locl helth office tht they hd hd symptoms of gstroenteritis. Exct dte nd time of symptom onset for these people were not known. For one of the five, lbortory dignosis confirmed infection with S. Enteritidis. Descriptive sttistics Among the 91 respondents, 52 mtched the cse definition: 22 hd lbortory-confirmed infection nd symptoms of slmonellosis, one hd lbortory-confirmed infection without symptoms nd 29 hd symptoms nd no lbortory confirmtion. A further three of the respondents hd symptoms of slmonellosis with lbortory confirmtion, but s the time of onset ws outside the period defined in the cse definition they were excluded from the nlysis. All further nlysis therefore reltes to 88 respondents, of whom 52 were clssified s cses nd 36 s non-cses (ttck rte: 59% (95% CI: 48 70)). Of the 52 cses, 35 were femle nd 17 were mle. The ttck rte ws lower mong the mle (17/37; 46% (95% CI: 30 63)) thn the femle wedding guests (35/51; 69% (95% CI: 54 81)) (chi-squre test p = 0.04). Informtion on ge ws unvilble for two of the 88 respondents: the medin ge of the remining 86 respondents ws 39 yers (rnge: 2 70). Tble 1 shows the distribution of cses nd ttck rte by ge group. The mjority of cses hd dirrhoe (49/52), stomch crmps (43/52), hedche (36/52), joint nd muscle che (32/52) nd nuse (29/52). Fever nd Tble 1 Cses of slmonellosis fter wedding prty nd ttck rte by ge group, Bvri, Germny, summer 2010 (n=50) of Age group of cses Attck rte respondents (yers) (n = 50) (n = 86) % (95% CI) (1 81) (3 100) (46 77) (35 71) (30 93) Age ws unknown for two cses mong the 88 wedding guests who responded to the questionnire. 3

4 vomiting ws experienced by 28 nd 21 of the 52 cses, respectively. Figure Cses of slmonellosis fter wedding prty by dy nd time of symptom onset, Bvri, Germny, summer 2010 (n=51) of cses Dy nd time of onset known Dy of onset known, time of onset unknown Dy 0 Dy 1 Dy 2 Dy 3 Dy 4 Dy nd time of symptom onset Dy 0 = dy of the wedding prty. One cse with lbortory-confirmed infection but without symptoms ws excluded. A totl of 10 cses with medin ge of 36 yers (rnge: 23 71) were hospitlised between 2 nd 10 dys (totl number of hospitl dys: 55) due to gstroenteritis. The distribution of cses by dy nd time of symptom onset is shown in the Figure. When dy of onset ws known, but time of onset ws unknown (n=3), the cses were systemticlly ttributed to the time period of for the pproprite dte.) Anlyticl sttistics In univrible nlysis, the exposure to 18 out of 63 individul dishes ws ssocited with significntly incresed reltive risk of slmonellosis (Tble 2). Only the following three items were ssocited with reltive risk >2: wedding rings (pst, rice, crrots nd myonnise), prwn sld (prwns, tomtoes nd myonnise) nd Greek noodles (rice noodles nd tomtoes). When the dishes were grouped ccording to the four different courses served t the wedding prty (soup, lunch buffet, cke, dinner buffet), only exposure to the lunch buffet ws significntly ssocited with infection (Fisher s exct test p<0.01). As there were no cses Tble 2 Univrible nlysis of dishes served t wedding prty, consumption of which ws significntly ssocited with onset of slmonellosis, Bvri, Germny, summer 2010 Dish exposed Exposed of cses Attck rte s percentge not exposed Not exposed of cses Attck rte s percentge Reltive risk (95% CI) Chi-squre test p vlue Lunch buffet Wedding rings (pst, rice, crrots, myonnise) ( ) < Glss noodle sld ( ) < Prwn sld ( ) < Lyered sld with herring ( ) Greek noodles (rice noodles nd tomtoes) ( ) Stuffed mushrooms ( ) Fried ubergines ( ) Cheese sld with myonnise nd grlic ( ) Stuffed tomtoes ( ) Cesr sld ( ) Peppers ( ) Olives ( ) Broccoli-culiflower cheese ( ) Grlic mushrooms ( ) Grilled turkey fillet with crem suce ( ) Dinner buffet Tomtoes nd mozzrell ( ) Shrimps with crem ( ) Cold trout with slices of lemon ( ) p<0.05; 95% CI did not include

5 mong the people who did not prtke of the lunch buffet, reltive risk of 4.49 (95% CI: 0.82 ) for exposure to dishes of the lunch buffet ws estimted by exct Poisson regression. When ll dishes were grouped ccording to common ingredients nd type of preprtion, two groups of dishes served s prt of the lunch buffet nd four groups of dishes served t the dinner buffet were ssocited with significntly incresed reltive risk of infection (Tble 3). Notbly the group of dishes served t lunch tht contined myonnise ws ssocited with the highest reltive risk of 7.83 (95% CI: ). Exposure to t lest one dish of the group of dishes contining myonnise could ccount for 44 out of the 46 cses who could be unmbiguously clssified s exposed or non-exposed (for six cses nd two non-cses, unmbiguous clssifiction ws not possible s they were uncertin bout their consumption of one dish in this group). Consumption of n incresing number of dishes contining myonnise ws ssocited with n incresing risk of infection (Tble 4); this trend ws sttisticlly significnt (Cochrn-Armitge test for trend p<0.001). The difference in reltive risk of infection between mle nd femle prticipnts ws no longer significnt fter djustment for the consumption of lunch dishes contining myonnise (dt not shown). In the multivrible nlysis, the following three groups of dishes remined significntly ssocited with infection nd were retined in the finl model: the group of dishes tht contined myonnise served t lunch, those contining myonnise served t dinner nd the suces served t lunch (Tble 5). Environmentl investigtion Seven people from the ctering compny were present t the wedding prty: one dditionl person ws involved only in the food preprtion t the cterer s fcilities. Of the seven t the wedding prty, one developed symptoms of slmonellosis on the return journey fter the event nd further five becme ill the following dy. Only the driver did not hve ny symptoms of slmonellosis. All seven stted tht they hd eten vrious dishes from the buffet s well s different ckes provided by wedding guests. According to the cterer, only the met dishes were prepred t the cterer s fcilities in dvnce. All suces, ntipsti nd slds were mde from commercilly produced ingredients nd were prepred t the wedding prty venue. Food items tht required cooling Tble 3 Univrible nlysis of groups of dishes served t wedding prty, consumption of which ws significntly ssocited with onset of slmonellosis, Bvri, Germny, summer 2010 Group of dishes exposed Exposed of cses Attck rte s percentge not exposed Not exposed of cses Attck rte s percentge Reltive risk (95% CI) Chi-squre test p vlue Lunch: dishes contining myonnise ( ) <0.001 Lunch: suces ( ) Dinner: dishes contining myonnise ( ) Dinner: vegetbles ( ) Dinner: shellfish ( ) Dinner: other sefood ( ) p<0.05; 95% CI did not include 1. Tble 4 Assocition between the number of dishes contining myonnise consumed per respondent nd the risk of slmonellosis fter wedding prty, Bvri, Germny, summer 2010 of dishes contining myonnise consumed per respondent of respondents (n=80) of cses (n=46) Reltive risk (95% CI) Wld test p vlue Reference ( ) ( ) ( ) > ( ) <0.001 Eight respondents, for whom clssifiction ws not possible s they were uncertin bout their consumption of one dish in this group, were excluded from the nlysis. 5

6 were trnsported in cooling boxes nd stored in cooling units t the venue. At the venue, the lunch buffet ws set up for six hours, between 2 p.m. (the buffet ws set up one hour before it ws vilble to guests) nd 8 p.m. No checks were crried out of holding tempertures of the wrm or cold dishes. Slds nd other cold dishes were not cooled during this period. Photogrphs of the buffet showed tht number of the cold dishes were decorted with non-edible flowers (such s roses), which were inserted into the food. The flowers hd been purchsed t wholesle retiler. Ckes supplied by number of wedding guests were stored without cooling until they were served t 20:00. The dishes of the dinner buffet were not cooled. They were first served t 10 p.m. It is uncler for how long the dishes of the dinner buffet were served; however, it is known tht the ctering personnel deprted t Inspection of the ctering fcilities nd interview of the ctering stff reveled number of shortcomings contrvening Europen food hygiene regultions. The fcilities used by the cterer were not registered with the locl uthorities. There were no records of the required stff trining on food hygiene. No temperture controls of cooling devices or trnsport boxes were crried out, nor were tempertures monitored during preprtion or serving of wrm dishes. There were no records of HACCP concept plnning or implementtion. The compny ws bnned from ctering until proof of chnges in their prctices hd been provided to the locl uthorities. Inspection of the kitchen fcilities t the wedding prty venue did not identify ny deficiencies. No food smples could be secured from either the venue or the cterer. However, four food smples, ech consisting of mixture of dishes, were provided for lbortory nlysis by wedding guests who hd tken home leftovers from the buffet. Tble 5 Multivrible nlysis of groups of dishes served t wedding prty, consumption of which ws significntly ssocited with onset of slmonellosis, Bvri, Germny, summer 2010 Reltive risk Vrible (95% CI) Lunch: dishes contining myonnise Wld test p vlue 0 1 Reference ( ) Dinner: dishes contining myonnise 0 1 Reference ( ) Lunch: suces 0 1 Reference ( ) p<0.05; 95% CI did not include 1. Lbortory results A totl of 38 of the respondents stted in their questionnire tht they hd submitted stool smple for lbortory nlysis vi their GP or t hospitl. Slmonell species were isolted from stool smples from 26 of them. For 24 of these isoltes, S. Enteritidis ws identified s the serovr. Stool smples from ll eight of the ctering personnel were positive for Slmonell. Swbs from both the wedding prty venue nd the ctering compny s kitchen fcilities were negtive. In two of the four food smples tested (one contined severl met dishes served t the lunch buffet, the other contined severl sefood dishes served t the dinner buffet), S. Enteritidis ws isolted in stndrd 25 g of smple mteril. At the Ntionl Reference Centre, eight isoltes from the respondents, two isoltes from ctering personnel nd two isoltes from the food smples were chrcterised. All 12 isoltes were identified s S. Enteritidis, phge type 4/6. Moleculr subtyping of these S. Enteritidis strins by PstI-SphI ribotyping showed tht ll strins possessed n indistinguishble ribotype pttern: ribotype 3 ccording to the Ntionl Reference Centre rbitrry designtion of ribotypes of S. Enteritidis PT 4/6. (To dte, the Ntionl Reference Centre hs identified 37 different ribotype ptterns mong isoltes of phge type 4/6 (dt not shown) Discussion This pper describes the investigtion of n outbrek of S. Enteritidis PT4/6 mongst the guests of wedding prty. The cohort study showed tht vriety of dishes were ssocited with significntly incresed risk of infection: in prticulr consumption of group of lunch dishes contining myonnise ws ssocited with high reltive risk. Despite the constrint of two-week dely between the wedding prty nd the questionnires being sent out, prticipnts ppered to recollect well which dishes they hd consumed. All respondents were contcted by the locl helth office by telephone within four dys of the wedding, which ment tht both cses nd non-cses were mde eqully wre of the outbrek close to the event, reducing the risk of recll bis. The isoltion of S. Enteritidis from two of the food smples t the wedding prty ws judged to show tht the food served posed helth risk, s ll the food items were redy for consumption without requiring further preprtion or heting. The isoltion of indistinguishble Slmonell strins from the food smples s well s from stool smples of respondents nd ctering personnel supported the hypothesis tht the outbrek ws food-borne. Besides the use of clssicl methods (serotyping nd phge typing), moleculr methods such s PstI-SphI ribotyping re essentil for identifying genomic polymorphism, in order to differentite 6

7 between individul strins of the highly clonl serovr Enteritidis nd thus confirm common source of the pthogen. There re severl possibilities for the source of the Slmonell contmintion in this outbrek. Myonnise is well-recognised vehicle of contmintion when rw egg is used s n ingredient [8-12]. However, in this outbrek ll cold dishes nd slds were mde from commercilly prepred ingredients. As commercilly produced myonnise nd suces re conventionlly bsed on psteurised ingredients, it is unlikely tht they would be the primry source of contmintion. Commercil myonnise by itself is lso not suitble for Slmonell propgtion, due to its low ph djusted by cetic cid. However, ddition of myonnise to other sld ingredients my lter the overll cetic cid concentrtion of the mixture, thus providing suitble bse for prolifertion once the pthogen hs been introduced by secondry contmintion [13]. None of the ctering stff sid they hd symptoms of slmonellosis before the wedding prty. However, shedding of pthogens in the prodroml phse of infection, s well s symptomtic excretion of Slmonell, hs been reported in the literture [14,15]. Thus the possibility of food contmintion by n symptomtic shedder mong the ctering personnel cnnot be excluded. Contct during storge or trnsport between rw ingredients tht my hve been contminted (such s mets or fish) nd dishes redy for consumption my hve been nother possible source of cross contmintion. Interestingly, number of the cold dishes nd slds (including the wedding rings ) nd number of other dishes lso contining myonnise served t the lunch nd dinner buffets hd been decorted with flowers. Europen food sfety lws stte tht food contmintion during primry production must be minimised by controlling potentil contmintion in soil, wter nd fertiliser. Cultivtion of non-edible flowers, however, is not subject to these lws nd flowers re therefore t incresed risk of pthogen contmintion. The flower decortion in vriety of cold dishes nd slds my hve served s the source of secondry Slmonell contmintion in this outbrek. Unfortuntely, it ws not possible to test this hypothesis s only some of the dishes could be clerly identified in photogrphs of the buffets nd there were no other records of which dishes hd been decorted with flowers nor ws it possible to obtin this informtion from the ctering stff. Unvilbility of systemtic smples of ll the dishes, or ingredients, which could hve helped to elucidte the source of contmintion, posed further limittion to the outbrek investigtion. The environmentl investigtion reveled number of infringements of food sfety regultions, including lck of stff trining nd the bsence of records of food sfety concept ccording to the HACCP principles. Lck of temperture controls for food storge nd trnsport s well s prolonged presenttion of buffet dishes t room temperture provided idel conditions for pthogen prolifertion, regrdless of the primry source of contmintion. While they do not replce officil controls, the HACCP principles re centrl to the Europen concept of food sfety by helping food business opertors to ttin high stndrd of food sfety. Successful implementtion of procedures bsed on the HACCP principles requires the full coopertion nd commitment of food business employees. Adequte trining of personnel is centrl to chieving this gol [2]. The outbrek investigted in this study demonstrtes the consequences of lck of stff trining nd the filure to identify hzrds to food sfety, s well s filure to implement control mesures to mitigte such hzrds. The use of flowers s food decortion demonstrted insufficient understnding of the potentil for contmintion through products tht re not intended for food production nd therefore not subject to food hygiene regultions. Intelligently implemented food hygiene concepts not only benefit the consumer but re lso very much in the interest of the food business opertor, whose business cn be thretened by food-borne outbreks. Initil hygiene nd food sfety trining for food business opertors should therefore lso explin microbiologicl principles underlying food sfety prctices in order to equip the businesses with the required bckground knowledge nd motivtion to design nd implement n intelligent food sfety/haccp concept, including the considertion nd identifiction of potentil sources of contmintion. Explicit mention of the dngers of the use of non-edible flowers for decortion should be considered in guides to good prctice, which re vluble instrument to id food business opertors with complince with food hygiene rules nd with the ppliction of the HACCP principles. Acknowledgments The uthors would like to thnk ll their collegues t the locl helth offices nd locl veterinry offices in Ansbch, Kssel nd Werr-Meissner for their help nd collbortion in the outbrek investigtion. We thnk S. Kulbe nd D. Busse for their skilful ssistnce in phge typing nd B. Knüppel for ribotyping. The uthors would lso like to thnk M. Dehnert, I. Krginnis nd K. Alpers t the Robert Koch Institute for their dvice throughout the investigtion. References 1. Europen Centre for Disese Prevention nd Control (ECDC). Annul epidemiologicl report on communicble diseses in Europe Stockholm: ECDC; Regultion (EC) No 852/2004 of the Europen Prliment nd of the Council of 29 April 2004 on the hygiene of foodstuffs. Officil Journl of the Europen Union. Luxembourg: Publictions Office of the Europen Union :L 139. Avilble from: LexUriServ.do?uri=OJ:L:2004:139:0001:0054:EN:PDF 3. 7 Meldepflichtige Nchweise von Krnkheitserregern des Gesetzes zur Verhütung und Bekämpfung 7

8 von Infektionskrnkheiten beim Menschen (Infektionsschutzgesetz- IfSG). [ 7 Notifible lbortory dignosis of pthogens of the Germn lw for the prevention nd control of infectious diseses in humns]. Lw for the prevention nd control of infectious diseses in humns]. Ministry of Justicein collbortion with Juris GMBH; first relesed 20 Jul 2000, lst mended 28 Jul Germn. Avilble from: bundesrecht/ifsg/gesmt.pdf 4. Wrd LR, de S JD, Rowe B. A phge-typing scheme for Slmonell enteritidis. Epidemiol Infect. 1987;99(2): László VG, Csórián ES, Pászti J. Phge types nd epidemiologicl significnce of Slmonell enteritidis strins in Hungry between 1976 nd Act Microbiol Hung. 1985;32(4): Liebn E, Grci-Migur L, Gurd-Petter J, McDowell SW, Rnkin S, Opitz HM, et l. Slmonell enteric serotype Enteritidis phge types 4, 7, 6, 8, 13, 29 nd 34: comprtive nlysis of genomic fingerprints from geogrphiclly distnt isoltes. J Appl Microbiol. 2002;92(2): Rbsch W, Prger R, Brun P, Methner U. Slmonell in poultry flocks nd humns--s. enteric subspecies enteric serovr Enteritidis in the pst. Berl Munch Tierärztl Wochenschr. 2007;120(7-8): Food poisoning ssocited with the consumption of myonnise. CDR (Lond Engl Wkly). 1991;1(33): Jrdine A, Ressler KA, Bothm S, Irwin MJ, Shdbolt C, Vlly H, et l. An outbrek of Slmonell typhimurium 108/170 t privtely ctered brbeque t Sydney sports club. Foodborne Pthog Dis. 2011;8(11): Mson BW, Willims N, Slmon RL, Lewis A, Price J, Johnston KM, et l. Outbrek of Slmonell Indin ssocited with egg myonnise sndwiches t n cute NHS hospitl. Commun Dis Public Helth. 2001;4(4): dos Sntos SM, Kupek E. Seril outbreks of food-borne disese in Blumenu, Brzil, cused by Slmonell Enteritidis. Brz J Infect Dis. 2000;4(6): Orteg-Benito JM, Lngridge P. Outbrek of food poisoning due to Slmonell typhimurium DT4 in myonnise. Public Helth. 1992;106(3): Erickson MC. Microbil risks ssocited with cbbge, crrots, celery, onions, nd deli slds mde with these produce items. Compr Rev Food Sci Food Sf. 2010;9(6): Avilble from: doi/ /j x/pdf 14. Todd EC, Greig JD, Brtleson CA, Michels BS. Outbreks where food workers hve been implicted in the spred of foodborne disese. Prt 5. Sources of contmintion nd pthogen excretion from infected persons. J Food Prot. 2008;71(12): Zu einnem noskomilen Ausbruch durch S. Enteritidis in Fuld. [A nosocomil outbrek of S. Enteritidis in Fuld]. Epidemiologisches Bulletin. 2007;48: Germn. Avilble from: Infekt/EpidBull/Archiv/2007/48 07,templteId=rw,property =publictionfile.pdf/48_07.pdf 8