eggs and salmonella food -poisonin an evalñuation

8/2/2019 eggs and salmonella food -poisonin an evaluation

1/8

J. Med. Microbiol.- ol. 34 (1991),65-720 1991ThePathological Societyof GreatBritain and IrelandREVIEW ARTICLEEggs and salmonella food-poisoning :an evaluationJ. P. DUGUID and R. A. E. NORTH*University of Dundee, DD 1 4HN and Departmentof Hospitality Management, Leeds Polytechnic, Leeds LS1 3HE

Summary. Evidence on the extent of the part played by infected hens eggs in causingsalmonella food-poisoning is inconclusive. The role of freshly cooked shell eggs is currentlymuch exaggerated. Prevention should be sought through improved catering practices andkitchen hygiene, and attempts to eradicate salmonellas from laying flocks are likely to beineffective.

IntroductionFrom early this century it has been known that

along with many other sources of salmonella, eggs aresometimes infected and sometimes cause food-poison-ing in man. * Ducks eggs were mainly implicatedand hens eggs only rarely, their role being thoughtsmaller than that of meat, pork, poultry meat andunpasteurised milk.

During and after the Second World War, outbreakswere traced to imported spray-dried and bulk liquidegg, in which salmonellas from a few contaminatedeggs had multiplied and disseminated during prepa-ration and storage. Pasteurisation required by theLiquid Egg (Pasteurization) Regulations 1963 hasmade such products relatively safe.3

Most of the recent egg-associated outbreaks havebeen due to products such as mayonnaise, ice-creamand cold desserts, in which salmonella can multiplyprofusely and which are eaten without cooking afterthe addition of raw egg. In these nutritious products afew bacteria from an eggs shell or contents canmultiply to huge numbers during storage for 1 or 2days at warm room temperature. With a generationtime of 80 min at 20C, one bacterium can become abillion (lo9) in 40 h, and with a generation time of 40min at 25C, it can do so in 20 h.

In contrast, shell eggs boiled, fried, poached orscrambled in the home have until recently beenthought to cause infection only rarely, despite theknowledge that light cooking which leaves some yolkor albumen liquid does not kill all contained bacteria,e.g., in duck eggs boiled for less than 8 min.2Reasons for the relative safety of freshly cookedhens eggs are seen in recent findings that very feweggs from infected flocks contain salmonellas and thatthe number of bacteria in these infected eggs is usuallyso low as to be below the human oral infective dose.Received 12 July 1990; ccepted 6Aug. 1990.

The general restriction of human (non-typhoid)salmonellosis to the food-poisoning mode of infectionindicates that profuse multiplication of the bacteria ina foodstuff is needed to yield a dose sufficiently largeto infect by mouth.

Recent concernEvidence regarding the increase of food-poisoning

in Britain in 1987 and 1988 has been interpreted bysome epidemiologists as showing that hens eggs havenow become a common source of food-poisoning. Itled the junior health minister, Mrs Currie, to state on3 December 1988 that most of the egg production ofthis country, sadly, is now infected with salmonella,which wrongly phrased pronouncement exaggeratedthe danger from eggs and aroused much public anxietyand political concern.

The Agriculture Committee of the House of Com-mons then examined evidence on the role of eggspresented by officers of the Department of Health(DOH), Ministry of Agriculture, Fisheries and Food(MAFF), Public Health Laboratory Service (PHLS)and other bodies, and published a First Report inFebruary 1989.49 Whilst admitting the evidence thateggs had become the commonest source of salmonellaoutbreaks was inconclusive, the Committee concludedthat the Government should not wait for cast-ironevidence before taking action and that it was right tomake egg production the first focus of its attention(Report, p.

In this situation of doubt, the Chief Medical Officerof the DOH justifiably advised the public that foodscontaining raw egg should be avoided, that for thevery young, very old, pregnant and sick, eggs shouldbe thoroughly cooked (yolk hard), but for other persons(the great majority) there was very little risk fromeating eggs cooked however preferred.

But there has been controversy over the Govern-65


2/8

66 J . P. D U GU ID A N D R. A. E. NORTHments hasty introduction of a policy for compulsorybacteriological testing of laying flocks (on 16March1989 under the Testing of Poultry Flocks Order 1989)and the slaughter of flocks found to contain anyinfected birds (Zoonoses Order 1989, effective from 1March 1989).This policy is out of line with practice in most otherdeveloped countries and is damaging to the Britishegg industry, particularly the small producers. It wasimposed without sufficient knowledge of the epi-zootiology of salmonellosis in laying flocks to givereason for believing it would be practicable andeffective.

The Ministry, moreover, implemented the slaughterpolicy almost immediately after it had advised theAgriculture Committee in a joint MAFF/DOH mem-orandum on 1 1 January 1989 (Report, p. 3) that:Salmonella organisms are present in many areas ofthe environment and are persistent. There is notechnically sound method which is certain to eliminatethem from laying houses. Even after very thoroughcleansing and disinfection, salmonella may recur.Because of this, a policy of slaughtering flocks withthe aim of eradicating salmonella would not have arealistic prospect of success. There would be a seriousrisk that even after an infected flock had beenslaughtered out and the premises disinfected, subse-quent flocks in the same building could becomeinfected from the environment.

Salmonella enteritids PT4The basis of the belief that hens eggs have become

a major source of salmonella food-poisoning n Britainwas the emergence since 1985 of one particular strain,Salmonella enteritidis phage type (PT) 4 as the cause ofa large proportion of human cases. In PHLS evidenceto the Agriculture Committee it was reported that inthe first 10 months of 1988 in England and Wales, 36%of 384 recorded outbreaks of salmonella food-poison-ing (Report, p. 17) and 46% of23 038 bacteriologicallyproven cases (Report, p. 6) were due to the PT4strain.

This strain has been found in flocks of laying hensand broiler (poultry meat) chickens and it has beenassumed that these poultry are its major source forman. Of 108 isolates of PT4 from food animalsexamined by the PHLS in 1987, 81% were frompoultry, the others from turkeys, ducks, cattle, pigsand sheep (Report, p. 8). MAFF reports showed thatovert infections with S . enteritidis were uncommon incattle and pigs, but became common in chickens,mainly broilers, in 1987 and 1988 (Report, p. 21).

The assumption that poultry are the main source ofthe PT4 strain may well be correct, but it is a weaknessof the case against eggs and poultry meat that nothorough investigation has been made of the strainsoccurrence in symptomless intestinal carriers indifferent food animals, pets, vermin, wild birds andkitchen staff, for it is such inapparent, carrier

infections that are most likely to be sources of food-poisoning.

Whilst herds and flocks may contain only a few(e.g., 1%) of symptomless carriers, cross-contamina-tion with their intestinal contents during slaughterand processing may infect the carcases of manyoriginally uninfected animals. Such cross-contamina-tion presumably explains why in PHLS surveys in1987 about 60% of retailed chicken carcases containedsalmonella (Report, p. 17).

Eggs versus poultry meatIf poultry are now the major source of salmonella

food-poisoning, the relative importance of eggs andpoultry meat remains in question and is relevant tothe Governments decision to impose a slaughterpolicy on laying, but not broiler flocks.Explaining this decision on 24 October 1989, theParliamentary Secretary to the MAFF stated that theGovernment had targeted eggs as the main health riskbecause they are eaten raw and part-cooked, while itis generally accepted that thorough cooking killssalmonella in chicken. This confidence in an under-standing of the need for thorough cooking of chickenseems misplaced in view of the many outbreaks offood-poisoning traced to its consumption. It alsoignores the great potential for cross-contaminationfrom raw carcases.

Until 1988 poultry meat was considered the majorsource of salmonella food-poisoning in Britain6* ndthe evidence that it has now been overtaken by eggs isinconclusive. In PHLS evidence to the AgricultureCommittee (Report, pp. 17, 18) it was noted that inonly a minority (under 20%) of outbreaks was a foodvehicle presumptively identified. Among these, in1987 only six outbreaks were associated with egg-containing foods but 41 with poultry meat. In the first10 months of 1988, by contrast, 46 outbreaks wereassociated with eggs and 28 with poultry meat.

The PHLS later cited the figures for 1988 and 1989shown in the table.* They indicate an increase in theproportion of outbreaks associated with eggs in theseyears. This increase may partly reflect the effect of thenew interest in eggs stimulating a more frequent

Table. Numb er of recorded outbreaks of sa lmon-ella food-poisoning in E ngland and W ales andnumbers of outbreaks associated with eggs andpoultry meatEgg- Poultrymeat- All

outbreaks outbreaks OutbreaksYear associated associated

1987 6 41 42 11988 34 27 4551989 42 24 955Total 82 92 1831


3/8

EGGS AND SALMONELLA 67reporting of outbreaks in which an egg source couldbe suspected. There are other reasons, moreover, whythe PHLS figures probably overestimate the role ofeggs and underestimate that of poultry meat.

Egg-containing ood sIt should not be assumed that in most or all

outbreaks traced to an egg-containing food the foodreceived its salmonella from an egg. In some or manysuch outbreaks the food may have received thesalmonella by cross-contamination via hands orutensils from another source in the kitchen, e.g., anuncooked infected poultry carcase, human carrier,food ingredient contaminated by vermin, or vegetablescontaminated by wild birds.

In some well investigated outbreaks, the role of eggshas been strongly indicated by a failure to detect anyother source and a finding of the epidemic strain inthe flock of hens known to have supplied the eggs. Butin most supposedly egg-associated outbreaks thepublished information has been fragmentary and theevidence incriminating eggs as the source has beenincomplete or absent.

In evidence to the Agriculture Committee (Report,pp. 256-290) one of us (R.A.E.N.) analysed thefindings in 13 outbreaks reported as egg-associated forwhich adequate information could be obtained andfound that in only two of them was the evidencesufficient to demonstrate that eggs might have beenthe primary source.In a MAFF/DOH memorandum to the AgricultureCommittee (Report, p. 1 9 ) , 30 outbreaks attributedto eggs were reported for the first three-quarters of1989, but in only 10 were the reasons implicating eggsreported to the PHLS Communicable Disease Surveil-lance Centre (CDSC): in three a statistical associationand in four the finding of salmonella in the supplyingflock.

When evidence of the association is so oftenincomplete or absent, conclusions drawn from thenumber of outbreaks reported as egg-associated mustbe viewed with suspicion.

Poultry meat and cross-contaminationThe role of poultry meat is not confined to the many

outbreaks traced to its cons~mption,-~or infectedcarcases also act as potent sources of cross-contami-nation to other foods, which then may cause outbreaks.

Chicken carcases are almost universally present inthe kitchens of hotels, restaurants, hospitals and othercatering establishments and it is known that manycarcases contain salmonella, e.g., 60% of retailedchickens (Report, p. 17). The number of bacteria ona carcase may be very large. From work in Hudders-field Polytechnic it has been reported that the dripfrom a frozen carcase during defrosting may containbillions of bacteria (J. Mottishaw, personal commu-nication). Traces of contaminated fluid on hands and

utensils may seed a variety of foods in the kitchen andeven introduce as many bacteria as may, withoutfurther multiplication, constitute an infective dose.

The role of cross-contamination in causing a givenoutbreak is difficult to detect and in most instances islikely to be unrecognised and unrecorded. That it is asignificant and not merely hypothetical danger hasbeen shown in some thoroughly investigated cases.Thus, food-poisoning with S . enteritidis PT8 affected14 guests at a wedding lunch who ate cold roast pork,but none of 45 guests who ate hot roast turkey. Pigswere not the source of the salmonella. The cold cookedpork had been contaminated in the kitchen fromuncooked turkey carcases which were later found tocontain the PT8 strain and to have been obtainedfrom a flock infected with that strain.

A major factor facilitating cross-contamination isthe design of commerical kitchens. Traditionally suchkitchens were designed for what was called thepartie system, with different kinds of foods preparedby chefs working in separate areas, from which theproduct was supplied on a raw-in, cooked-out basis.Over the years the parties have contracted and workwith both raw and cooked meats and other foods isconducted in a general partie, where cross-contam-ination is almost inevitable.

Food-poisoning is commonly attributed to thefailure of food handlers to wash their hands betweencontacts with raw and cooked foods. Our experienceof food premises is that too often wash basins aresoiled, obstructed or lacking in soap and dryingfacilities, and where adequate are frequently unused.During a food hygiene conference in Swindon in 1988a survey in the washroom revealed that 65% ofparticipants using the facilities did not wash theirhands.Sporadic cases

Only a minority, probably less than lo%,of cases ofsalmonella food-poisoning occur in recognised out-breaks affecting several persons who have shared acommon meal. The majority are sporadic, in singlepatients not known to have shared a foodstuff withother sufferers. The true number of sporadic cases isunknown, for most are unreported (Report, p. 249).Whereas there were 23 038 bacteriologically provencases of salmonella food-poisoning in England andWales in the first 10 months of 1988 (Report, p. x4),the total number of cases may have been 10 times asmany.

From evidence suggesting that 46 of 384 outbreaksof salmonella food-poisoning in 1988 (38 of 191 S.enteritidis outbreaks) were egg-associated (Report, p.17) it has been suggested that many or most sporadiccases are caused by eating infected shell eggs, boiled,fried, poached or scrambled, but cooked too lightly tokill salmonella. As few eggs are infected and rarely


4/8

68 J. P. DUGUID A N D R . A. E. NORTHmore than one person eats from the same egg, mostcases socaused would be sporadic.

It is unjustifiable thus to extrapolate to sporadiccases from uncertain conclusions about outbreakscaused by egg-containing foods.. In these foods a fewbacteria initially present often have the opportunity,before consumption, to multiply profusely to yield alarge infective dose. But when whole eggs are eatenshortly after breaking from the shell, there is noopportunity for such multiplication.

The degree of risk in eating shell eggs depends onthe proportion of retailed eggs containing salmonella,the number of bacteria usually present in the infectedeggs, the extent of bacterial survival on cooking, andthe number of bacteria that have to be swallowed inorder to cause illness, i.e., the size of the oral infectivedose.Frequency of infection in eggs

Salmonella is found on the shell of eggs, from soilingwith hen faeces, more often than in the contents. Somebacteria from the shell may enter egg-containing foodsand there multiply to infective levels, but it is only thebacteria present in the eggs contents that can pose adanger to persons eating freshly cooked shell eggs.

Widely differing frequencies of internal infection ineggs have been reported. In small, recently infectedflocks the proportion of eggs containing salmonellamay be large, e.g., 5 out of 10 eggs from a smalldomestic fl o ~ k , ~0 of 613 eggs from an artificiallyinfected flock of 42 hens,14 and 11 of 1119 eggs fromtwo naturally infected flocks of 23 and 15 birds; eachinfected egg contained less than 10 S . enteritidis PT4bacteria.

In large flocks associated with food-poisoning theproportion of internally infected eggs has been smaller,about 0.1% or less: e.g., S . enteritidis PT4 in 4 of 2000eggs (Report, p. 205),0 of and 5of 17 000,17andS. typhimurium in 3 of 1137 eggs* and 0 of 1000.19

As most laying flocks are uninfected, the overallfrequency of infection in British eggs is likely to bemuch lower than 1 in 1000. A MAFF study in 1989showed a contamination rate of about 1 in 15000eggs(cited by Stevens et al.) and even lower rates havebeen estimated.

Britons eat on average just under three eggs a week,or about 150 a year (Report, p. 11). If 1 in 15 000 eggsis infected, the chance of a person eating an infectedegg is very low, only 1 in 100 per year. In a populationof nearly 60 million, about 600 000 persons would eatan infected egg each year, but in most such cases toofew bacteria would be present in the egg to constitutean infective dose.

Numberof bacteria in egg contentsAlthough a matter of crucial importance, few studies

have been made of the number of salmonella bacteria

in the contents of intact, naturally infected eggs. Thosereported suggest that the number per egg is usuallyvery small, in the order of 1-100. Each of 11 infectedeggs from two flocks contained fewer than 10 S.enteritidisPT4 bacteria, though some of these eggs hadbeen stored for 5 days at 20C before examination.After inoculation of S . pullorurn into the ovary, thenumber of bacteria recovered from the eggs was onlyfrom 0.9 to 4*6/g.o

The probable reason for the presence of only smallnumbers of bacteria is the protective effect of the eggscomplex system of membrane barriers and anti-bacterial components in the 22 Thesedefences are overcome when eggs are subjected toartificial inoculation or broken with mixing of yolkand albumen, and they deteriorate on prolongedstorage under warm conditions.

The contents of the egg may become infected in twoways: (a) through the shell, where it is laid on dampearth or litter soiled with infected droppings, and (b)transovarially, into the forming yolk or albumen fromthe ovary or oviduct of a systemically infected hen.

Infection through the shellThis mode of infection has been studied experimen-

tally. In one eggs at 35C were immersed in acold suspension of S. montevideo, S . oranienburgor S .typhimurium bacteria and sampled after drying andstorageat 29C for periods of up to 29 days. Salmonellawas absent from the yolk and albumen immediatelyafter inoculation, penetrated in small numbers intosome eggs during the first week, but penetrated intoand multiplied to large numbers (107-109/ml) n mosteggs during the third and fourth weeks. The eggsdefences were therefore highly effective in the firstfortnight of storage at 29C, but later broke down.

The relevance of such findings to the risk fromtrans-shell infection under natural conditions is un-clear. Eggs laid on damp, faeces-soiled earth will drawin not only salmonella, but also numerous putrefactivefaecal bacteria such as pseudomonas and gram-negative anaerobes. Some of these saprophytes willmultiply faster than salmonella at ambient tempera-tures and probably make the egg inedible before itssalmonella content is large.

Transovarian infectionDuring the eggs formation, before it is covered by

the shell, the yolk or albumen may become infectedfrom the ovary or oviduct in a recently infected,septicaemic hen. 9 14* 24 The eggs contents are thusinfected with salmonella without an associated, warn-ing infection with spoilage bacteria, and the yolk maybe infected before it is protected by the anti-bacterialalbumen.

The yolk may have some yet unrecognised defences,such as its relatively anaerobic state and large contentof antib~dies,~hich may explain why only small


5/8

EGGS AND SALMONELLA 69numbers of bacteria have been found in the yolks ofeggs naturally infected by the transovarian route. 5Possibly, too, it is nutritionally deficient for the non-proteolytic, non-lipolytic salmonellae which requirefree amino acids, of which there are no knownquantities in the undisrupted yolk.26Some workers have shown rapid multiplication ofsalmonella after artificial inoculation into the yolk. Inone study2' an inoculum of about 10 cultured bacteriamultiplied to about lo9 in 24 h at 23C and 10" in48 h. But such inoculation introduces growth mediumcomponents and causes some mixing of yolk withalbumen, changes that appear to disrupt the egg'sdefences. These experimentalfindings cannot be takento represent what commonly happens in naturallyinfected, intact eggs, otherwise huge, instead of onlysmall numbers of bacteria would be found in sucheggs.Although profuse multiplication does not take placein the generality of transovarially infected eggs, itpossibly does so in some abnormally formed ormistreated ones. Such exceptional multiplication mayexplain those rare family outbreaks caused by a dishprepared from shell eggs only a few hours beforeconsumption.28 n some cases, however, an apparentabsence of sufficient time for bacterial multiplicationafter the breaking-out of eggs may be due to wronglyremembered or reported circumstances.

Survival on cookingWhilst salmonella bacteria in free egg fluid are

killed within a few minutes when heated at atemperature as low as 60"C, light cooking does notreliably rid them from whole eggs because heatpenetrates only slowly through the static mass ofviscous contents.

Boiling an egg at 100C will kill salmonella on theshell, but many studies have shown that the cookingof boiled or fried eggs to yield a firm white but partlysoft yolk does not kill all bacteria in the yolk. In testson eggs into the yolks of which S. enteritidis had beeninjected, viable bacteria were recovered from thoseboiled for only 4 min or fried on only one side (yolksoft), though not from those boiled for over 8 min orfried on both sides (yolk hard).27 Cooking must becontinued until both white and yolk are completelycoagulated to ensure freedom from viable bacteria.

Size of oral infective doseThe food-poisoning (non-typhoid) types of salmon-

ella such as S. enteritidis and S. typhimurium causeboth overt and symptomless intestinal infections in awide range of domestic and wild animals and birds.They must commonly contaminate the environment,including earth, water and vegetables, and frequently

enter the human mouth in small numbers on contam-inated hands, utensils, uncooked vegetables, anddrinking and bathing water. Yet clinical infection inman is largely confined to the food-poisoning mode.

This limitation is generally, and probably correctly,attributed to an inability of small numbers (ems., -1000) of swallowed bacteria to overcome the defencesof the alimentary canal and a need for preliminarymultiplication in a nutritious foodstuff to yield a largeenough oral dose (e.g., over lo5 bacteria) to overcomethese defences.

The defences include the bactericidal action of theusually strongly acid (pH


6/8

70 J. P.DUGUID AND R. A. E. NORTHSmall-dose infection

Though large doses of salmonella are generallyrequired for infection, there is evidence that inexceptional circumstances a small dose (1-1000 bac-teria) may suffice. Thus in rare outbreaks caused bychocolate and cheese, only small numbers of bacteriacould be recovered from the causal foodstuff.32

Only 2.5 S. eastbourne bacteria/g were found inchocolate balls infecting 119 and theinfective dose was estimated to be about 50 S. napolibacteria in chocolate bars infecting245 and1-6 S. typhimurium bacteria in cheddar cheese infect-ing 1500 persons. 9

It has been suggested that dehydrated bacteriaentrained in the lipids of fatty foodstuffs may beprotected from killing by the gastric acid, though whya few such protected bacteria should be able toovercome the intestinal defences is unclear. Possiblycomponents of chocolate and cheese help the bacteriacolonise the throat before being swallowed. Butwhatever the explanation, the findings of McCulloughand Eisele3* 6 from the feeding of salmonellas in eggnog show that the lipids of egg do not have the sameenhancing effect as those of chocolate and cheese.

Hospital infectionsEvidence for small dose infection with food-

poisoning salmonellas has been seen in certain hospitaloutbreaks among neonates, infants, debilitated adultsand old persons not caused by infected food.40*1 Theinfection appeared to be spread from patient to patientand from nurse to patient by vehicles such as hands,clothing, towels, wash basins and ward dust, unlikelyto deliver more than a few bacteria into the mouth. Ina general hospital where 102 persons developedgastroenteritis and 150 became symptomless carriers,the causal strain of S . typhimurium was found in warddust, on sheets used by excretors and in the sputum offive patients.41

It has been suggested that defective gastric andintestinal defences in the young, debilitated and oldmake such subjects specially vulnerable to small-doseinfection. In neonates, for instance, gastric acidity islow in the first weeks of life42 and the protectiveintestinal flora takes a few days to become properlyestablished.

Though some small-dose infections may be due todefective defences, others may be caused by the entryof salmonella through specially susceptible non-oralroutes such as the conjunctiva and respiratory tract.The bacteria may first multiply profusely in these sitesbefore seeding in large numbers into the hostileenvironment of the stomach and intestine. There aremany opportunities in hospital for bacteria to enterthe eye or nasopharynx, as when a nurses contami-nated hand washes a babys face, when adult patientswash their faces in contaminated basins or whenpatients inhale contaminated ward dust.

The great susceptibility of the conjunctival routewas demonstrated in guinea-pigs by Moore;43 infec-tions resulted from the dropping of 100 S . enteritidisbacteria into the eye, but not from the feeding of lo8bacteria. And the conjunctival route has been foundto be a 100-fold more susceptible than the oral routefor infection of mice with S . t y p h i m ~ r i u m . ~ ~ompa-rable observations in man are lacking, but a volunteerinto whose nasal antrum 25 salmonella bacteria wereinoculated, developed acute sinusitis in 4 h andgastroenteritis in 37h, showing the promoting effectof initial multiplication in the upper respiratory

Case-control studiesof sporadic infectionsThe bacteriological evidence of the small proportion

of eggs infected and the small numbers of bacteriapresent suggests that not many sporadic infections arelikely to be due to the eating of undercooked shelleggs, but this conclusion needs to be confirmed byepidemiological studies. Unfortunately, only a fewsmall studies of sporadic infections have been re-p ~ r t e d , ~ ~ ~7 insufficient for firmconclusions.

food histories were obtainedfrom 160 patients with sporadic S . enteritidis PT4infections and up to three uninfected neighbours(controls) per patient. The illness was found to besignificantlyassociated with the consumption of shop-bought sandwiches containing mayonnaise(p=0.00004) and bought pre-cooked chicken(p= 0.006),foods liable to cross-contaminationduringpreparation and to unrefrigerated storage beforeconsumption.

But there was only a weak association (p=O.O2)with the eating of lightly cooked fried, poached orscrambled eggs, or omelette, and no significantassociation with the eating of soft-boiled eggs. It isdifficult, therefore, to understand how the authors ofthe study concluded that fresh shell eggs . . . arevehicles of S . enteritidis phage type 4 infection inindigenous sporadic cases. Larger studies are re-quired, but to our minds the present results suggestthere is little danger from eating freshly cooked shelleggs.

In the larger

Preventive measuresAs salmonella and other food-poisoning bacteria

have common, almost ubiquitous sources in man,animals and the environment, their absence fromfoods as supplied from the farm cannot be ensured byGovernment action. The correct sphere for govern-mental control is in the regulation of practices incommercial food-processing and catering to minimiseopportunities for cross-contamination and bacterialmultiplication.

This view accords with advice in the Code ofPractice for the Control of Salmonellae n Commercial


7/8

EGGS A ND SALMONELLA 71Laying Flocks issued in 1988 by the MAFF and BritishPoultry Federati~n,~~hich stated : Salmonellaorganisms are widespread and their complete elimi-nation from the environment can never be expected.There is evidence that an increase in food poisoninghas been associated with use of raw foodstuffs ofanimal origin in the preparation of food for humanconsumption. This problem could be largely overcomeby applying good kitchen hygiene and practices . . ..

Special attention should be given to the preparationand handling of those foodstuffs that have caused alarge proportion of outbreaks in recent years, namelymayonnaise, cold desserts and pre-cooked savouries.

MayonnaiseMayonnaise was the probable or possible vehicle of

infection in 14 of the 46 egg-associated outbreaks ofsalmonella food-poisoning reported to the CDSC forthe first 10 months of 1988 (Report, pp. 31-345). Theapparently rising incidence of food-poisoning frommayonnaise made in the home or a catering establish-ment has been attributed to reduction in the bacteri-cidal vinegar content to meet consumer preference fora blander pr ~d uc t. ~ uring unrefrigerated storagethe new product allows the multiplication of bacteriaderived either from its content of rawegg or by cross-contamination from another source in the kitchen.

Home-made mayonnaise should be acidified withvinegar to a pH between 3-6 and 4-0.49Commercialproducts should be made with pasteurised egg and, toprevent their later contamination in catering establish-ments, should be supplied in sealed containers not solarge as to encourage re-use on a second day afteropening. In a PHLS report on a mayonnaise outbreakin the House of Lords the need for guidelines on thepreparation of mayonnaise in commercial premiseswas suggested. Such guidelines issued in 1955 inDenmark appear to have been effective, though oneoutbreak in a day nursery from home-made mayon-naise was reported in 1980.50

Avoidance of cross-contaminationGreater emphasis should be given to the prevention

of cross-contamination in catering kitchens, as egg-based foods are not made safe by excluding contamin-ation from eggs. They may be cross-contaminatedfrom another source in the kitchen, so that such cross-contamination and the opportunity for later multipli-cation of bacteria must be prevented. Raw poultrycarcases are probably the commonest source ofcontaminationand the most useful preventive measuremight be a requirement that catering establishmentsshould obtain poultry only from sources, such asirradiation plants, which could guarantee their free-dom from salmonella.

Slaughter policyThe present requirement for the testing of laying

flocks and the slaughter of those containing anyinfected birds is at variance with the advice given bythe MAFF to the Agriculture Committee (Report, p.3,5 quoted above), and will probably not significantlyreduce the amount of food-poisoning.

Periodic (12-weekly) testing of flocks by culture offaeces or cloaca1 swabs collected by the flock keeperwill not reliably detect infection either soon enoughafter its entry into a flock or in flocks where only aminority of birds are carriers.

Buildings cannot be completely disinfected after aflock has been slaughtered and the replacement flocksare liable to be re-infected by residual salmonellas inthe building, infected feeding stuff (Report, p. xviii4),the droppings of mice, which are impossible to excludefrom poultry houses, human carriers, or salmonellasfrom the droppings of wild birds carried into thepoultry house on the keepers feet. Re-infection hasalready been reported in the replacement flocks onsome of the farms where flocks were slaughtered in1989, and we know of five such re-infected farms,housing a total of 385 000 birds.

In the first year of implementation of the policy,from March 1989 to February 1990, over a millionhens were slaughtered, but the amount of salmonellafood-poisoning continued to increase. In December1989 the Agriculture Committee took encouragementfrom reports that the increase in the number of humaninfections with S . enteritidisPT4 was not as great in1989 as in 1988 (Report, p. viig). It is known thatparticular salmonella strains spontaneously becomemore and then less common over periods of years, andno doubt the PT4 strain will eventually decline, butthat decline should not be taken to indicate success ofthe slaughter policy.

A detailed account and analysis of the data,inquiries, reports and pressures leading to impositionof the policy has recently been p ~ b l i s h e d . ~ ~

It must be questioned whether the problem ofsalmonella in eggs deserves the great attention it hasrecently received. Salmonella food-poisoning is a non-contagious disease and most cases are scarcely moreserious than the common cold and similar upperrespiratory viral infections, which are nearly a thou-sand times commoner. Only a minute proportion offood-poisoning cases are fatal and the fatalities aremainly in old debilitated persons. There are manygraver health problems to be faced.

References1 . Scott WM. Food po isoning due to eggs. Br Med J 1930:2: 56-

59.2. Leading article. Eggs and salmonella nfection s.Br Med J 1944;2 :760-76 1.3. Richmond M, Arbuthnott JP, Baird-Parker AC et al . Th emicrobiological safety of food. London, HMSO. 1990; I:118.


8/8

72 J. P. DUGUID AND R. A. E. NORTH4. House of Commons Agriculture Committee. First Report,Salmonella n Eggs. Vol I, Report and Proceedings of theCommittee, J. Wiggin (Chairman) 1989. London, HMSO.5. House of Commons Agriculture Committee. First Report,Salmonella in Eggs. Vol 11, Minutes of Evidence andAppendices,28 February 1989.London, HMSO.6. Humphrey TJ, Mead GC, Rowe B. Poultry meat as a source of

human salmonellosis in England and Wales. EpidemiolInfect 1988; 100: 175-184.7. Reilly WJ, Forbes GI, Sharp JCM, Oboegbulem SI, CollierPW, Paterson GM. Poultry-borne salmonellosis in Scot-land.Epidemw l Infect 1988; 101: 115-122.8. Public Health Laboratory Service and State Veterinary Service.Update on salmonella nfection (edn2). Colindale, London,PHLS/SVS, January 1990.9. House of Commons Agriculture Committee. Salmonella inEggs-A Progress Report, 20 December 1989. London,HMSO.10. Small RG, Halliday GJ, Haxton AR, Hird MR, Sharp JCM,Wallace JM. Salmonella food-poisoning from turkeys.Health Bull (Edinb)1976; 34: 206-21 1.11. Cruickshank JG, Humphrey TJ. The carrier food-handler andnon-typhoid salmonellosis. Epidemiol Infect 1987; 8: 223-230.12. MacAuslan E. Clean behind the scenes.Caterer and Hotelkeeper5 April 1990, Letters: 24.13. Paul J, Batchelor B. Salmonella enteritidis phage type 4 andhenseggs.Lancet 1988; 2: 1421.

14. Timoney JF, Shivaprasad HL, Baker RC, Rowe B. Eggtransmission after infection of hens with Salmonellaenteritidisphage type 4. Vet Rec 1989; 25 :600-60 1.15. Humphrey TJ, Baskerville A, Mawer S, Rowe B, Hopper S.Salmonella enteritidis phage type 4 from the contents ofintact eggs: a study involving naturally infected hens.Epidemiol Infect 1989; 103: 415-423.16. Stevens A, Joseph C, Bruce J et al. A large outbreak ofSalmonella enteritidis phage type 4 associated with eggsfrom overseas.Epidemiol Infect 1989; 03 :425-433.17. North RAE. Failure of government: a report on the risk of foodpoisoning from eggs and egg products. Leeds PolytechnicDepartment of Hospitality Management 1989:27-28.18. Philbrook FR, MacCreadyRA,Van Roekel H et al. Salmonel-losis spread by a dietary supplement of avian source. NEngl JMe d 1960; 263: 713-718.19. Chapman PA, Rhodes P, RylandsW . Salmonella typhimuriumphage type 141 infections in Sheffield during 1984 and1985: association with hens eggs. Epidemiol Infect 1988;20. Forsyth RH, Ross WJ, Ayres JC. Salmonella recovery followinggastro-intestinal and ovarian inoculation in the domesticfowl.Poult Sci 1967;46: 849-855.21. Board RG, Fuller R. Non-specific antimicrobial defences ofthe avian egg, embryo and neonate. Bwl Rev 1974; 49: 15-49.22. Tranter HS, Board RG. The antimicrobial defence of avianeggs; biological perspective and chemical basis. J Appl

Bwchem 1982; 4: 295-338.23. Stokes JL, Osborne WW, Bayne HG. Penetration and growthof Salmonella in shell eggs. Food Res 1956; 21: 510-518.24. Hopper SA, Mawer S. Salmonella enteritidis in a commerciallayer flock. Vet Rec 1988; 123: 51.25. Dadrast H, Hesketh R, Taylor DJ. Egg yolk antibody detectionin identification of salmonella infected poultry. Vet Rec1990; 126: 219.26. Griffin HD, Perry MM, Gilbert AB. Yolk formation. In:Freeman BM (ed) Physiology and biochemistry of thedomestic fowl, vol5. London, Academic Press. 1984: 345-380.27. Humphrey TJ, Greenwood M, Gilbert RJ, Rowe B, ChapmanPA. The survival of salmonellas n shell eggs cooked under

101:75-82.

simulated domestic conditions. Epidemw l Infect 1989; 103:35-45.28. Devine E, Finnie BG. Outbreak of egg-borne salmonella food-poisoning. Commun Dis Scotland W kly Rep 1988; 11:5.29. GiannellaRA,Broitman SA, ZamcheckN. Gastric acid barrierto ingested micro-organisms n man : studies in vim and invitro. Gut 1972; 13 : 251-256.30. Dixon JMS. The fate of bacteria in the small intestine.J Path01Bacterwll960; 79: 131-140.31. Meynell GG. Antibacterial mechanisms of the mouse gut. 11.The role of Eh and volatile fatty acids in the normal gut.Brit J Exp Pat h1 1963; 44: 209-219.32. Blaser MJ, Newman LS.A review of human salmonellosis: I.Infective dose. Rev Infect Dis 1982; 4: 1096-1 106.33. Naylor GRE. Incubation period and other features of food-borne and water-borne outbreaks of typhoid fever inrelation to pathogenesis and genetics of resistance. Lancet1983; 1: 864-866.34. Boring JR, Martin WT, Elliott LM, Isolation of Salmonellatyphi-muriumfrom municipal water, Riverside, California,1965. Am J Epidemw ll971; 93 : 49-54,35. McCullough NB, Eisele CW. Experimental human salmonel-losis. I. Pathogenicity of strains of Salmonella meleagridisand Salmonella anatum obtained from spray-dried whole

egg. J Infect Dis 1951; 8:278-289.36. McCullough NB, Eisele CW. Experimental human salmonel-losis. 111. Pathogenicity of strains of Salmonella newport,Salmonella derby and Salmonella bareilly obtained fromspray-dried whole egg.J Infect Di s 1951 89: 209-21 3.37. Craven PC, Mackel DC, Baine WBet al. International outbreakof Salmonella eastbourne infection traced to contaminatedchocolate.Lancet 1975; 1:788-793.38. GillON,Sockett PN, BartlettC L R e t a l .Outbreakof Salmonellanapoli infection caused by contaminated chocolate bars.Lancet 1983; 1:574-577.39. DAoust J-Y. Infective dose of Salmonella typhimurium incheddar cheese.Am J Epidemwl 1985; 122: 717-720.40. Bate JG, James U. Salmonella yphimurium infectiondust-bornein a childrens ward. Lancet 1958; 2: 713-715.41. Datta N, Pridie RB , Anderson ES. An outbreak of infectionwith Salmonella typhimurium in a general hospital. J Hyg

42. Miller RA. Observations on the gastric acidity during the firstmonth of life. Arch Dis Child 1941;16 : 22-30.43. Moore B. Observationspointing to the conjunctiva as the portalof entry in salmonella infections of guinea-pigs. J Hyg1957; 55 :414-433.44.Duguid JP, Darekar MR, Wheater DWF. Fimbriae andinfectivity in Salmonella typhimuriu m. J Me d Microbw l1976; 9: 459-473.45. MacKenzie CR, LivingstoneDJ. Salmonellae n fish and food.

SAfric Med J 1968; 42: 999-1003.46. Coyle EF, Palmer SR, Ribiero CD et al. Salmonella enteritidisphage type 4 infection:association with hens eggs. Lancet47. Cowden JM, Lynch D, Joseph CA et al. Case-control study ofinfections with Salmonella enteritidis phage type 4 in

England.BY Med J 1989; 299: 771-773.48. Ministry of Agriculture, Fisheries and Food and British PoultryFederation. Code of practice for the control of salmonellaein commercial aying flocks. London, MAFF/BPF. 1988.49. Perales I, Garcia MI. The influence of pH and temperature onthe behaviour of Salmonella enteritidis phage type 4 inhome-made mayonnaise.Lett Appl Microbioll990; 10 : 19-22.50. Rosdahl N. Salmonella enteritidis in gastroenteritis in a daynursery. (Epidemiologicalobservations in a primary food-borne infection). Ugeskr-Laegar 1980:142 :2795-2799.51. North R, Gorman T (eds). Chickengate: independent analysisof the salmonella in eggs scare. London, IEA Health andWelfare Unit, Institute of Economic Affairs.1990.

1960; 8 :229-241.

1988; 2: 1295-1297.

Documents

eggs and salmonella food -poisonin an evalñuation