International Journal of Food Microbiology 43 (1998) 15

Type E botulism associated with vacuum-packaged hot-smoked whitesh

H. Korkeala a , *, G. Stengel b , E. Hyytia a , B. Vogelsang b , A. Bohl c , H. Wihlman d , P. Pakkala e , S. Hielm a
a

Department of Food and Environmental Hygiene, University of Helsinki, P.O. Box 57, FIN-00014 Helsinki, Finland b Food and Veterinary Institute Schleswig-Holstein, Neumunster, Germany c Veterinary and Food Ofce, Itzehoe, Germany d Department of Meat Hygiene, National Veterinary and Food Research Institute, Helsinki, Finland e National Food Administration, Helsinki, Finland Received 6 March 1998; accepted 27 May 1998

Abstract On January 16, 1997 two Germans got botulism after eating hot-smoked Canadian whitesh produced in Finland. The serum sample of one of the patients contained 6 MLD/ ml of botulinum toxin. The type of toxin was identied as E by the toxin neutralization test and the botulinum neurotoxin type E (BoNT / E) gene was also amplied from the serum by polymerase chain reaction (PCR), but C. botulinum could not be isolated from the positive serum sample. The remains of the hot-smoked whitesh eaten by the patients contained botulinum toxin detected by the mouse bioassay and the BoNT / E gene as determined by PCR. C. botulinum was isolated from the sh sample and it was conrmed to be type E by the mouse bioassay and by PCR. Eleven other sh samples from the same lot did not contain botulinum toxin nor any BoNT gene. The incriminated food was processed on the 9th and 10th of January, 1997 from frozen whitesh imported to Finland from Canada. The pulsed-eld gel electrophoretic pattern of the isolated C. botulinum strain resembled a reference strain of North American origin. It did not match any C. botulinum strains isolated from the Baltic sea-bottom or from the sh caught in the area indicating that the sh was contaminated by C. botulinum in Canada. The conditions resulting in toxin production could not be identied. The safety problems associated with vacuum-packaged hot-smoked sh seem to be of utmost concern and the product is one of the most important botulism food vehicles processed on an industrial scale. Temperature monitoring and the use of timetemperature indicators are to be recommended in order to ensure adequate storage temperature from processing through to consumption. Allowing the use of nitrate and nitrite together with sufciently high NaCl concentration in this particular product should also be considered. 1998 Elsevier Science B.V. All rights reserved. Keywords: Clostridium botulinum type E; Botulism; Outbreak; Fish, hot-smoking; Vacuum-package

1. Introduction In recent years an average of 450 Clostridium

*Corresponding author.

botulinum outbreaks have been reported annually in the international literature; 12% of the outbreaks were caused by type E (Hatheway, 1995). C. botulinum type E is an aquatic organism (Huss, 1980) and both bottom sediments and sh are often

0168-1605 / 98 / $19.00 1998 Published by Elsevier Science B.V. All rights reserved. PII: S0168-1605( 98 )00080-4

H. Korkeala et al. / International Journal of Food Microbiology 43 (1998) 1 5

contaminated by this organism (Hielm et al., 1998b; Hyytia et al., 1998). Foodborne botulism is seldom contracted nowadays from commercial products and approximately 90% of cases are caused by home-prepared or homepreserved foods (Hatheway, 1995). Vacuum-packaging of hot-smoked whitesh has been shown to cause type E botulism in human beings (Koenig et al., 1964). The consumption of industrially processed sh products including vacuum-packaged products has increased in recent years. This can lead to C. botulinum type E outbreaks caused by commercial products. In Sweden two C. botulinum type E outbreaks associated with vacuum- packaged hotsmoked rainbow trout have been recorded recently (Anonymous, 1991; Oberg, 1994). More information on this kind of outbreak is needed in order to improve the safety of sh products. On January 16, 1997 two Germans took ill after eating hot-smoked whitesh manufactured in Finland. An investigation was immediately launched to nd the cause of the outbreak.

2.3. Laboratory methods

Clostridium botulinum was detected by a PCR method (Hielm et al., 1996) and the bacterium was isolated by combining the PCR method with that of the NCFA (1991). The toxicity of the clinical and food samples and the ability of the isolated strains to produce botulinum toxin were determined by the mouse bioassay (Kautter et al., 1992). The aerobic plate count was determined by the method of NCFA (1986) using plate count agar (Merck, Darmstadt, Germany). Characterization of the strains genome was performed by macrorestriction analysis and pulsed-eld gel electrophoresis (PFGE). In situ DNA isolation was performed using the method of Maslow et al. (1993), modied by formaldehyde xation of cells prior to lysis (Hielm et al., 1998a). Twelve strains of mostly North American origin (Hielm et al., 1998a) and 42 isolates from Finnish trout farms were used as C. botulinum type E reference strains.

3. Results 2. Materials and methods

3.1. Patients
The husband and wife and their grandchild ate hot-smoked whitesh on the 15th of January. The sh was not heated before eating. The 39 year-old woman ate about two thirds of the sh (approximately 600 g) and the 41 year-old man ate the rest. The taste of the sh was considered good by them. The two-year-old grandchild had a bone stuck in its throat and nished eating after two small pieces of sh. Both the woman and the man developed botulism but the child did not. The woman experienced symptoms of dizziness, blurred vision, dyspnea, headache, and dysarthria without diarrhea, and the incubation period was 7.5 h. The incubation period in the man was 13.5 h and his symptoms were nausea, abdominal cramps, headache, dizziness, vomiting, and diarrhea. The wife was placed on respiratory support on the 16th and 17th of January. On the 30th of January both patients were discharged from the hospital. The serum sample of the wife contained 6 MLD/ ml of botulinum toxin as determined by the mouse

2.1. Investigations of patients

The three family members two of whom contracted botulism type E, one of whom remained unaffected were interviewed as to the type and quantity of foods consumed, storage of foods, type and time of onset of symptoms. Samples of serum and gastric content taken from the two patients were studied.

2.2. Food investigation

Samples of part of the hot-smoked vacuum-packaged whitesh consumed by the aficted persons were studied, as were 11 sh from the same lot. Information on raw materials, methods of processing, and transport of the products was obtained by interviewing the persons in charge and by studying their HACCP records.

H. Korkeala et al. / International Journal of Food Microbiology 43 (1998) 1 5

bioassay. The type of the toxin was identied as E by the toxin neutralization test and the botulinum neurotoxin type E (BoNT / E) gene was also amplied from the serum by PCR but we were not able to isolate C. botulinum from the positive serum sample. The serum sample of the husband was negative in both tests. Samples of the stomachs gastric uids from both patients were PCR negative.

3.2. Food and strain analyses

The remains of the hot-smoked whitesh eaten by the patients were brought to the laboratory. The label on the vacuum-package indicated the manufacturer and the batch. The sh contained botulinum toxin measured by the mouse bioassay and the BoNT / E gene as determined by PCR. C. botulinum was isolated from the sample and it was conrmed to be type E by the mouse bioassay and by PCR. The isolated strain and the control strain type E Beluga both produced toxin of 400 MLD/ ml during the 3 days enrichment in TPGY broth at 268C. Eleven other sh samples from the same lot did not contain botulinum toxin. Neither did they contain the BoNT / E gene. The samples were still PCRnegative after incubation of 1 and 3 weeks at 20 and 48C, respectively. Aerobic plate counts of the samples were under 200 cfu / g besides two samples with APCs of 700 cfu / g and 9.2 ? 10 4 cfu / g. Macrorestriction analysis of the isolated strain showed a closer resemblance to some North American C. botulinum group II reference strains, than to Finnish ones. Strain 31-2570 E, isolated in 1974 in the USA, shared eight restriction fragments out of thirteen with the outbreak strain. This gives a Dice similarity coefcient of 62%, as opposed to a coefcient of 36% between the outbreak strain and its closest relatives isolated outside North America.

ing during the night between the 9th and 10th of January. The sh was initially dehydrated at 408C for 30 min and then smoked in four steps (30 min at 608C, 30 min at 708C, 40 min at 758C and 40 min at 708C). After smoking, the sh was chilled to 2.58C for 90 min. According to the manufacturer the NaCl concentration of sh was 1.8% (w / w). The 452 kg batch was vacuum-packaged in single sh packages weighing about 600 g each and labelled in the early morning of the 10th of January. Truckloading of the batch was done on the same day between 11 and 12 a.m. A 352 kg amount of the whitesh was exported to Germany and the rest of the batch was delivered to the Finnish market. The temperature during transportation of the batch to Germany was recorded and was mostly between 1 and 38C, occasionally it rose to 58C. The batch arrived in Germany on the 12th of January at 8.30 a.m. and was retained in the storage hold of the importer. On the 13th of January at 3 a.m. the sh was loaded into a truck and was delivered to a retail shop at 8.30 a.m. The married couple bought the sh that same day. There are no documented temperature data, but the temperatures were checked later. They were 58C in the storage hold of the importer, 28C at the start of the truck delivery, 58C in the retail shop, and 48C in the refrigerator of the patients.

4. Discussion Clinical and laboratory ndings showed that the vacuum-packaged whitesh was the vehicle of the type E botulism outbreak. Due to other similar C. botulinum outbreaks in the 1990s (Anonymous, 1991; Oberg, 1994) we have to take into account the risks caused by C. botulinum type E associated with vacuum-packaged hot-smoked sh products. Vacuum-packaged hot-smoked sh is one of the most important botulism food vehicles processed on an industrial scale. The positive PCR result of the serum sample was unexpected. After incubation the TPGY tubes were not turbid, indicating sterility. DNA fragments that served as PCR templates for BoNT / E primers were apparently present in the serum. The possible use of PCR in studying serum samples and the signicance of the observation warrants further research.

3.3. Processing of sh and transport conditions

The incriminated food was processed on the 9th and 10th of January. The frozen whitesh was imported to Finland from Canada on the 20th of December. The sh was thawed in cold water for 3 h. Brining took place in a container for 10 h. The NaCl content of the brine was 9.5% and the temperature 568C. Smoking started immediately after brin-

H. Korkeala et al. / International Journal of Food Microbiology 43 (1998) 1 5

Obviously, the smoking temperature applied did not kill all C. botulinum spores present in the sh (Alderman et al., 1972). The isolation of a C. botulinum type E strain from the incriminated sh also supports this conclusion. The PFGE pattern of the isolated strain resembled reference strains of North American origin but not any of the C. botulinum strains isolated from the Baltic Sea bottom sediments or from the sh caught in the area. This indicates that the sh was already contaminated by C. botulinum in Canada and it was not a postprocess contamination. The NaCl concentration used was not high enough to prevent the outgrowth of C. botulinum type E spores. Thus the sh must have been at a sufciently high temperature for a sufciently long time after processing for the strain to produce botulinum toxin. According to the information given, the sh was smoked only 5 days before consumption. This indicates that the toxin production had been rapid or that there had been a marked temperature abuse during storage or transport of the sh. The toxin production capacity of the C. botulinum type E strain isolated from the incriminated sh did not differ from that of the C. botulinum type E Beluga strain. The data obtained did not reveal any marked temperature abuse during the transport and the good microbiological quality of the other sh of the same lot conrmed this. It could be possible that the information given on the processing, transport or storage was not quite correct. Another possibility is that the smoking temperature was not high enough to destroy possible botulinum toxin formed before processing. The conditions resulting in toxin production could not be identied. In the Swedish outbreaks the mistakes resulting in toxin production could not be pointed out either. Incorrect storage conditions were proposed (Anonymous, 1991). The temperature monitoring documentation and the use of timetemperature indicators should be recommended to ensure the adequate storage temperature from processing to consumption. Sufciently high NaCl concentration in the product would prevent the outgrowth of C. botulinum type E spores (Pelroy et al., 1982; Graham et al., 1996). Due to public health aspects food industry, however, tries to decrease the amount of NaCl in foods. Nitrate and nitrite are also known to possess inhibitory properties on botulinum toxin

formation in sh products (Pelroy et al., 1982; Hyytia et al., 1997a). However, due to the possible formation of carcinogenic nitrosamine compounds (Walters, 1992) the use of nitrite and nitrate is prohibited in many countries (Pelroy et al., 1982; European Parliament and the Commission of the European Communities, 1995). The consumption of vacuum-packaged hot-smoked sh products is associated with increased botulism risks and further research on the preventive methods is needed to improve the safety of the products.

Acknowledgements The authors wish to thank Dr. Kammerer, Medical University, Lubeck, Germany, for supplying the patient samples and Dr. Schoo for information about the importer. Ms. Kirsi Ristkari, Ms. Maria Stark and Ms. Birte Weitzel are thanked for their excellent technical assistance.

References
Alderman, G.G., King, G.J., Sugiyama, H., 1972. Factors in survival of Clostridium botulinum type E spores through the sh smoking process. J. Milk Food Technol. 35, 163166. Anonymous, 1991. Botulism. Epid-aktuellt, 14 (6) 9. European Parliament and the Commission of the European Communities, 1995. European Parliament and Commission Directive 95 / 2 / EC. Ofcial Journal of the European Communities 61, 140. Graham, A.F., Mason, D.R., Peck, M.W., 1996. Predictive model of the effect of temperature, pH and sodium chloride on growth from spores of non-proteolytic Clostridium botulinum. Int. J. Food Microbiol. 31, 6985. Hatheway, C.L., 1995. Botulism: The present status of the disease. Curr. Top. Microbiol. 195, 5575. Hielm, S., Hyytia, E., Ridell, J., Korkeala, H., 1996. Detection of Clostridium botulinum in sh and environmental samples using polymerase chain reaction. Int. J. Food Microbiol. 31, 357365. Hielm, S., Bjorkroth, J., Hyytia, E., Korkeala, H., 1998a. Genomic analysis of Clostridium botulinum group II by pulsed-eld gel electrophoresis. Appl. Environ. Microbiol. 64, 703708. Hielm, S., Hyytia, E., Andersin, A., Korkeala, H., 1998b. A high prevalence of Clostridium botulinum type E in Finnish freshwater and Baltic Sea sediment samples. J. Appl. Microbiol. 84, 133-137. Huss, H.H., 1980. Distribution of Clostridium botulinum. Appl. Environ. Microbiol. 39, 764769.

H. Korkeala et al. / International Journal of Food Microbiology 43 (1998) 1 5 Hyytia, E., Eerola, S., Hielm, S., Korkeala, H., 1997. Sodium nitrite and potassium nitrate in control of nonproteolytic Clostridium botulinum outgrowth and toxigenesis in vacuumpacked cold-smoked rainbow trout. Int. J. Food Microbiol. 37, 6372. Hyytia, E., Hielm, S., Korkeala, H., 1998. Prevalence of Clostridium botulinum type E in Finnish sh and shery products. Epidemiol. Infect. 120, 245250. Kautter, D.A., Solomon, H.M., Luke, D.E., Bernard, D.T., Mills, D.C., 1992. Clostridium botulinum and its toxins. In: Vanderzant, C., Splittstoesser, D.F. (Eds.), Compendium of Methods for the Microbiological Examination of Foods. American Public Health Association, 3rd edition, pp. 605621. Koenig, M.G., Anderson, S., Cardella, M.A., Rogers, D.E., 1964. Clinical and laboratory observations on type E botulism in man. Medicine 43, 517545. Maslow, J.N., Slutsky, A.M., Arbeit, R.D., 1993. Application of pulsed-eld electrophoresis to molecular epidemiology. In:

Persing, D.H., Smith, T.F., Tenover, F.C., White, T.J. (Eds.), Diagnostic Molecular Microbiology: Principles and Applications, ASM Press, Washington, DC, pp. 563572. NCFA, 1986. Aerobic microorganisms. Enumeration at 308C in meat and meat products. Method no. 86. Nordic Committee on Food Analysis, 2nd ed. Espoo. NCFA, 1991. Clostridium botulinum. Detection in foods and other test materials. Method no. 80. Nordic Committee on Food Analysis, 2nd ed. Espoo. Oberg, S., 1994. Botulism. Epid-aktuellt 17 (8), 5. Pelroy, G.A., Eklund, M., Paranjpye, R.N., Suzuki, E.M., Peterson, M.E., 1982. Inhibition of Clostridium botulinum types A and E toxin formation by sodium nitrite and sodium chloride in hot-process (smoked) salmon. J. Food Prot. 45, 833841. Walters, C.L., 1992. Reactions of nitrate and nitrite in foods with special reference to the determination of N-nitroso compounds. Food Add. Cont. 9, 441447.