Series of incidents of Listeria monocytogenes non-invasive febrile gastroenteritis involving ready-to-eat meats
Correspondence to: J.A. Hudson, Food Safety Programme, ESR Limited, Christchurch Science Centre, PO Box 29–181, Ilam, Christchurch, New Zealand (e-mail: email@example.com).
Aims: A series of cases and outbreaks of febrile noninvasive gastrointestinal disease involving 31 identified cases was investigated in terms of the numbers and types of Listeria monocytogenes present in the suspect foods (ready-to-eat meats) and clinical samples from cases.
Methods and Results: Foods and faecal samples involved in the incidents were tested for the presence and number of L. monocytogenes . Isolates were typed by macrorestriction analysis using pulsed-field gel electrophoresis. The foods contained high levels of L. monocytogenes , in one case 1·8 × 10 7 g −1 . Faecal samples contained L. monocytogenes for up to 15 d after the contaminated food was consumed. All isolates from the food and faecal samples were of serotype 1/2 and were indistinguishable from one another by macrorestriction typing.
Conclusions: It is likely that the meats were contaminated either during their manufacture after they had been cooked or by underprocessing. The long shelf lives on these products would have allowed the contaminating L. monocytogenes to grow to the high numbers measured in this study, causing food poisoning as described.
Significance and Impact of the Study: Outbreaks of febrile noninvasive listeriosis are relatively rare. This report adds ready-to-eat meats to the range of foods that have acted as vehicles for such outbreaks.
In recent years it has become clear that Listeria monocytogenes infection can cause febrile gastrointestinal disease in otherwise healthy people (Anonymous 1997). This is in contrast to invasive listeriosis, which more usually affects the young, old, pregnant and immunocompromised. Outbreaks of febrile noninvasive listeriosis have involved a number of food vehicles including smoked mussels (Misrachi et al. 1991), shrimps (Riedo et al. 1994), rice salad (Salamina et al. 1996), chocolate milk (Dalton et al. 1997), rainbow trout (Miettenen et al. 1999), corn and tuna salad (Aureli et al. 2000) and imitation crab meat (Farber et al. 2000). In these outbreaks large populations of L. monocytogenes are typically reported as being present in the foods concerned, ranging from 1·9 × 105 g−1 to 1·2 × 109 g−1, and in one instance the median number consumed was estimated to have been as high as 2·9 × 1011 person−1 (Dalton et al. 1997).
Outbreaks of febrile listerial gastroenteritis are characterized by fever and diarrhoea in patients with an incubation period of 11 h to 1 week after exposure. However cases may also experience ‘flu’-like symptoms in the absence of any gastrointestinal symptoms (Salamina et al. 1996). The most common symptoms are fever, diarrhoea, muscle/joint pain, and headache. Salamina et al. (1996) noted that a sore throat occurred in 72% of cases, but in two subsequent outbreak reports this symptom has occurred less frequently, at 3% (Dalton et al. 1997) and 11·8% (Aureli et al. 2000) in adults.
This report discusses a series of incidents of noninvasive febrile gastroenteritis, which was traced to the manufacturer of a nationally distributed range of ready-to-eat meats.
In mid February 2000 the Christchurch Public Health Laboratory received two food samples in conjunction with reports from Health Protection Officers (HPOs) in the South Island. The reports linked the consumption of corned silverside from the same manufacturer with illness (Table 1). These cases (1 and 2) became ill with febrile gastroenteritis 17 and 32 h after consuming the meat. At least two more South Island cases (3 and 4) showing similar symptoms (Tables 1 and 2) were notified soon after. Both had eaten ready-to-eat meat from the same manufacturer.
Table 1. Microbiological data obtained for samples associated with the outbreak
|Case 1||Left over corned beef *||>2·5 × 105||1/2||96/2|
|Case 2||Left over corned beef||>2·5 × 106||1/2||96/2|
|Faecal sample||7·5 × 103||1/2||96/2|
|Case 3||Luncheon ham||Detected <100||1/2||96/2|
|Case 4||Faeces (3 samples)||Detected||–||–|
|South Island outbreak (n=7)|
| ||Luncheon ham from retailer but not consumed in the common meal||Detected <100||1/2||96/2|
|Corned silverside from retailer but not leftover from the common meal||2·5 × 102||1/2||96/2|
|Faeces (2 samples)||5 × 102||1/2||96/2|
| || ||5 × 102||1/2||96/2|
|North Island outbreak (n=21)|
| ||Left over ham||1·8 × 107||1/2||96/2|
| ||Faeces (20 samples)||Detected||1/2||96/2|
|Unopened retail packages|
| ||Bulk corned beef (5 samples)||1·1 × 103||1/2||96/2|
| || ||2·2 × 103||1/2||96/2|
| || ||1·0 × 103||1/2||96/2|
| || ||1·7 × 104||1/2||96/2|
| || ||2·5 × 102||1/2||96/2|
| ||Roast pork||1·5 × 102||1/2||96/2|
Table 2. Summary of symptoms reported
|Fever||7 (77·8)||16 (76·2)|
|Muscle pains||8 (88·9)||NR|
|Headache||6 (66·6)||17 (81·0)|
|Diarrhoea||3 (33·3)||14 (66·7)|
|Sore behind eyes||2 (22·2)||NR*|
|Nausea||1 (11·1)||16 (76·2)|
|Vomiting||1 (11·1)|| 6 (28·6)|
|Stomach pains||NR||12 (57·1)|
Based on the results of the tests on the initial two samples (Table 1) the manufacturer recalled product from one of its plants. The plant ceased production while local HPOs and representatives from the Ministry of Health investigated to ensure that all contaminated products were identified and the source of the contamination eliminated.
On the day of the recall seven people in the South Island became ill after eating a meal that included corned silverside, and shortly afterwards another incident occurred in the North Island involving 21 cases among 24 people attending a common meal where ham produced by the above-mentioned manufacturer was served. Here a supermarket had failed to remove the recalled product from the shelves despite being aware of the recall. These were the last known cases in this series of incidents.
Information other than microbiological analyses was collected and supplied by the HPOs and none of the cases is known to have had a predisposing condition.
Detection and enumeration of L. monocytogenes in foods
Food samples (25 g) were homogenized by stomaching for 2 min in a Colworth 400 stomacher (A.J. Seward, London, UK) with 225 ml of Listeria Enrichment Broth, Oxford Formulation (LEB) (CM862, Oxoid, Basingstoke, Hampshire, UK). The food enrichments were incubated in LEB at 30°C for 4 h before addition of filter-sterilized stock solutions of nalidixic acid, acriflavine and cycloheximide to achieve final concentrations of 40, 10 and 50 mg l−1, respectively. Incubation was continued for a further 44 h at 30°C. At 24 and 48 h a loopful of the enrichments was streaked to PALCAM agar (Van Netten et al. 1989) and Listeria Selective Agar (LSA), Oxford Formulation (1·07004, Merck, Darmstadt, Germany), which were incubated at 35°C. These plates were examined at 24 and 48 h for colonies characteristic of Listeria. To ensure the purity of the isolates up to five typical colonies, where available, were further subcultured to Trypticase Soy Agar (1·05458, Merck) plus 0·6% (w/v) Yeast Extract, which was incubated for 18–24 h at 30°C.
Suspect colonies were Gram-stained and tested for catalase, oxidase and tumbling motility in wet mounts. Subcultures were made to Columbia Blood Agar (CM331, Oxoid) containing 5% (v/v) defibrinated sheep blood to test for haemolysis. Haemolytic colonies were further tested for their ability to ferment 0·5% (w/v) D-glucose, maltose, rhamnose, mannitol, and xylose. Further identification was performed using the API Listeria kit (Bio Merieux, Mercy L'Etoile, France).
To enumerate Listeria in foods a 1 in 10 dilution of the food was made in peptone diluent and injured cells were allowed to resuscitate by incubating the sample for 1 h at room temperature. Volumes (0·1 ml) of 10-fold dilutions of the suspension were then spread onto two plates each of LSA and PALCAM agars and incubated at 35°C for 48 h. Suspect colonies were counted and up to five suspect colonies were subcultured from each plate and identified as described earlier. These identifications were used to attribute approximate the proportions of different species that might be present.
Detection of Listeria in faecal samples
Faecal samples were plated directly to LSA and PALCAM agars or homogenized in peptone diluent, a dilution series was performed and 0·1 ml-volumes were spread plated as described earlier to obtain a count. Plate incubation and colony identification were as described earlier.
Typing of Listeria isolates
Macrorestriction analysis using pulsed field gel electrophoresis (PFGE) and serotyping were carried out as described by Brett et al. (1998). Pulsed field gel electrophoresis was performed using SmaI and ApaI as restriction enzymes and the restriction fragments were separated using a contour-clamped homogeneous electric field apparatus (CHEF mapper, Bio-Rad Laboratories, Hercules, California, CA). The profiles obtained after digestion with restriction enzymes were compared visually. Isolates were considered to be indistinguishable if the SmaI and ApaI profiles matched exactly. Isolates were serotyped using the Bacto Listeria O serotyping reagents (Difco Laboratories, Detroit, MI).
Results for quantitative and qualitative analyses of the food and faecal samples are shown in Table 1. All isolates were identified as L. monocytogenes. In the cases where the food samples tested were left over from incriminated meals the numbers of L. monocytogenes present were high. In all but one sample the count was >2·5 × 105 g−1, and in the North Island incident the level was 1·8 × 107 g−1. One sample of luncheon meat from case 3 contained L. monocytogenes that was indistinguishable from the other isolates, but the number detected was <100 g−1. In the South Island outbreak no left-over corned beef was available for analysis. However, corned beef purchased from the retailer who supplied the meat implicated in the outbreak had 2·5 × 102L. monocytogenes g−1. A sample of luncheon ham from the same retailer was positive (<100 L. monocytogenes g−1), but the luncheon was not eaten during the common meal.
Five samples of corned beef and one of roast pork produced by the manufacturer and purchased unopened from retail were all found to contain L. monocytogenes in moderate numbers (Table 1).
Faeces from case 2 contained 7·5 × 103L. monocytogenes g−1 10 d after consumption of the implicated food. Two out of three faecal samples from case 4 were positive for L. monocytogenes approximately 15 d after consumption. Two cases from the South Island incident had 500 L. monocytogenes g−1 in their faeces 7 d after consumption of the implicated meal. All 20 faecal samples collected in the North Island incident yielded L. monocytogenes. These samples included three from asymptomatic diners, and as the four diners who did not provide samples were symptomatic, all diners were actually infected. Of those consuming the ham 87·5% became symptomatic, although in one case the symptom was of only 1·5 h in duration and consisted solely of a headache.
All isolates from cases involved in the outbreak were of serotype 1/2 and were of an indistinguishable PFGE type, as were the isolates from the ham, luncheon meat and the corned silverside produced by the same manufacturer.
Symptoms and incubation periods
Data for 30 cases are presented in Table 2. The predominant symptoms suffered in the South Island cases were ‘flu’-like, including muscle pains, fever and headache, with symptoms more usually associated with gastrointestinal disease, such as vomiting and diarrhoea, much less frequently reported.
In the larger North Island outbreak the symptoms were somewhat different in that gastrointestinal illness was prominent.
Incubation periods ranged from 12 to 101 h, but in most cases the symptoms began to appear at approximately 24 h. In the North Island incident the duration of the disease was from 1·5 h to 168 h.
Investigations at the manufacturer's plant failed to reveal any one problem that was responsible for the contamination that resulted in the outbreak. However, a number of potential concerns were identified in the manufacturing process. It was considered likely that the opening of a bagged product after cooking and its subsequent re-packaging afforded an opportunity for cross-contamination to occur. No other handling, such as slicing, was involved. There was also some question as to the adequacy of the heat treatment applied to products.
Foods involved in the outbreak were produced over a period of at least 3 months, from December 1999 until the recall in February 2000. The implicated ham had been produced in December and so at least 2 months had elapsed between the time of manufacture and consumption. Packs of meat supplied to supermarkets that were intended to be sliced before retail sale had 90-d periods between manufacture and their ‘best before’ date. Corned beef implicated in the cases that came to light on the 11th and 15th February 2000 was manufactured on 31st December 1999. The process for estimating the shelf life of the product was found to be flawed, and the shelf life was subsequently reduced to 6 weeks.
A serotype 1/2 L. monocytogenes strain of a characteristic PFGE type was involved in the incidents described in this study. In previous years this subtype of L. monocytogenes had been isolated from 1/20 clinical cases of invasive listeriosis in 1991, 1/14 in 1992, 4/11 in 1994, 5/15 in 1995, 2/10 in 1996, 17/33 in 1997, 3/17 in 1998 and 5/18 in 1999 (unpublished data). As this subtype represented a minority of clinical isolates in previous years, except in 1997 when there was an outbreak of nonperinatal invasive listeriosis (Anonymous 1998), the fact that all isolates from the cases and suspect foods described in this study were of an indistinguishable macrorestriction type is significant. Given the high numbers of L. monocytogenes found, the consistency in isolation of an indistinguishable subtype of L. monocytogenes from incident-associated implicated foods and unopened packs from retail produced by the manufacturer, and the symptoms displayed, it can be stated with confidence that all cases were part of a series of incidents where ready-to-eat meat products were the vehicle.
In previous studies, an outbreak of ‘mild’ listeriosis was attributed to serotype 4b (Riedo et al. 1994) and an outbreak of gastroenteritis in a day care facility to serotype 4 (Heitmann et al. 1997). In contrast, an outbreak in Italy, where a contaminated rice salad was the vehicle, was caused by serotype 1/2b (Salamina et al. 1996), as was a US outbreak associated with chocolate milk (Dalton et al. 1997). There is therefore no apparent link between febrile gastroenteritis caused by L. monocytogenes and any one serotype.
The symptoms reported in this outbreak are somewhat unusual because of the low prevalence of gastrointestinal symptoms in the South Island cases. However, Salamina et al. (1996) reported that four of 18 cases involved in the outbreak described did not suffer from gastrointestinal symptoms. In the North Island incident, as a result of the consumption of contaminated ham, however, the symptoms were more like those described in previous reports. It is possible that the symptoms experienced are related to the number of L. monocytogenes cells ingested.
In the outbreak involving ham, 21 of 24 (87·5%) people consuming the food became ill with symptoms of febrile gastroenteritis (Table 2). Assuming approximately 100 g of ham was eaten by each person at the meal, then the dose ingested to produce this response was of the order of 109 cfu. In the outbreak described by Dalton et al. (1997) an attack rate of 75% was recorded where the median population consumed was estimated as being as high as 2·9 × 1011 cfu. In the other outbreaks it is difficult to estimate dose–responses as the portion sizes were not detailed or the number of cells present were not accurately known. However, of all of the other outbreaks, the lowest number in food that has been shown to cause febrile noninvasive listeriosis is 1·9 × 105 cfu g−1 (Miettinen et al. 1999), although the serving sizes were not detailed. In this incident all five people eating the contaminated fish became ill with gastroenteritis, nausea, abdominal cramps and diarrhoea. Therefore consumption of more than, perhaps, 107 cells appears to be sufficient to cause L. monocytogenes febrile gastroenteritis at a high infection rate in some circumstances. It is possible that foods contaminated with lower numbers of L. monocytogenes may also cause febrile noninvasive gastrointestinal disease, and because this organism is not routinely screened for in clinical laboratories many cases of noninvasive listeriosis may evade detection. In instances where the symptoms are similar to those described, L. monocytogenes should be considered by clinical laboratories as a potential aetiological agent.
The result for case 3, where <100 L. monocytogenes g−1 was isolated, reinforces the link between meat products from the manufacturer and the type of L. monocytogenes involved. However, whether the disease experienced was associated with consumption of the luncheon meat sampled, with the low number of L. monocytogenes present, is questionable, but the symptoms experienced (fever and abdominal pain) could have resulted from the consumption of L. monocytogenes. Testing for L. monocytogenes in the faeces of this case might have supplied information as to whether low numbers of the organism can cause the gastrointestinal form of the disease. The data for this case are, however, not presented in Table 2 as there is some doubt as to the aetiology of the disease.
While the initial level of contamination of the products is not known, the long shelf lives marked on these products most likely were of prime importance in allowing the population of L. monocytogenes to reach the levels that they did. If foods with long shelf lives are produced that allow the growth of L. monocytogenes then it is clear that great efforts must be made either to ensure that the product does not become contaminated before its final packaging or to reduce the shelf life so that accidental contamination does not result in the growth of the organism to high numbers. An alternative strategy would be to re-formulate products to retard greatly, or inhibit altogether, the growth of this organism.
We thank the many HPOs involved in the recognition and investigation of these incidents, and the Public Health Laboratory staff who carried out the detection and enumeration work. Tecklok Wong, Rosemary Whyte, Jenny Bishop and Michael Baker are thanked for their constructive criticism of the manuscript.