Abstract
A Bacillus cereus–related emetic outbreak was reported in a Belgian kindergarten. High levels of emetic B. cereus (>1.5E+07 colony-forming units/g) were detected in the food leftovers, and the presence of an emetic strain was confirmed in feces. Emetic toxin levels ranging up to 4.2 μg/g were also quantified in the leftovers by liquid chromatography coupled to tandem mass spectrometry (LC-MS2) analysis. Those levels, although moderate in comparison with earlier published intoxications, provoked profuse-vomiting episodes in 20 toddlers aged between 10 and 18 months. Few studies have focused on the levels of emetic toxin implicated in food intoxications. This publication emphasizes the importance of defining toxic doses of emetic toxin among high-risk population groups.
Introduction
A
Detection methods for cereulide include cytoxicity-based assays using cell cultures and the boar spermatozoa motility assay, but they do not provide accurate quantitative results, only approximate toxicity titers. The development of sensitive analytical methods such as liquid chromatography coupled to mass spectrometry (LC-MS) allows for more precise quantification of the cereulide concentration in food (Häggblom et al., 2002). This may help to determine toxic doses for vulnerable groups and to set relevant criteria regarding consumer safety.
In Belgium, from 2007 until 2012, two to eight foodborne outbreaks were reported on a yearly basis in which B. cereus was identified as the causative agent. During this period 8 of 26 B. cereus outbreaks were caused by emetic B. cereus representing 147 cases and 1 death.
In this report, we describe a foodborne outbreak affecting 20 toddlers aged between 10 and 18 months caused by the consumption of a homemade mashed rice–cucumber–chicory meal contaminated with critical levels of the B. cereus emetic toxin.
Methods
Epidemiological information and sampling
In Belgium, inspectors of the Federal Agency for Safety of the Food Chain (FASFC) are responsible for food investigations and sampling of foods, whereas medical doctors of the Belgian Communities are responsible for collecting human samples. This outbreak was reported to the Health Inspection Service by the director of the kindergarten. Epidemiological information such as age and symptoms, but also the timeline and circumstances of the outbreak, were gathered by local health inspectors and inspectors of FASFC from the director of the kindergarten and from the parents using a standard questionnaire. All collected information was transmitted to the Belgian National Reference Laboratory of Food-borne Outbreaks (NRL-FBO). A case was defined as anyone who consumed a lunch at the kindergarten and who presented with vomiting episodes within 1 h.
Microbiological analysis was performed at NRL-FBO on four food leftovers and fresh chicory collected from the refrigerator at the kindergarten, and on uncooked rice that was used to prepare the implicated food. A soiled cloth used to clean vomit from contaminated surfaces was also analyzed. Three stool samples and one vomit sample were taken from four different human cases (toddlers).
Laboratory investigation
Food and environmental samples
The enumeration of B. cereus and coagulase-positive Staphylococcus aureus was performed according to the ISO 7932:2004 (ISO 2004) and ISO 6888:2003 (ISO, 2003) standards, respectively. The soiled cloth was soaked in 100 mL Buffered Peptone Water, and 10 μL of the obtained liquid was streaked on Mannitol egg Yolk Polymixin (MYP) agar (Bio-Rad, Marnes-la-Coquette, France) as adapted from ISO 7932:2004.
Human samples
A loopful of each sample of human stool and vomit was streaked directly on MYP agar and Baird-Parker agar (Bio-Rad) in order to search for B. cereus and coagulase-positive Staphylococcus aureus. The presence of Norovirus in stool samples was analyzed by reverse transcription quantitative polymerase chain reaction (PCR) using the primers recommended in the ISO/TS 15216:2013 (ISO/TS 15216, 2013) standard.
PCR assays
A PCR that detects the presence of the B. cereus emetic toxin encoding gene was performed on food and human isolates (Ehling-Schulz et al., 2004).
Quantification of cereulide by LC-MS2 analysis: Levels of cereulide in food and vomit samples containing B. cereus were quantified by LC-MS2 analysis as described previously (Delbrassinne et al., 2012a). As cereulide is an emetic toxin and because of the limited amount of sample, its presence was not tested in stool samples.
Results
Epidemiological investigation
At the end of August 2012, an outbreak at a Belgian kindergarten was reported to the Health Inspection. A total of 20 cases out of 22 exposed children were identified during this investigation. Case-patients were between 10 months and 1.5 years old. All suffered from vomiting episodes within 30 minutes following the lunch, indicating a point-source outbreak. None of the children had fever or diarrhea and none of them needed hospitalization. Cases all recovered within a few hours. Fever was reported in three children 2–3 days prior to the outbreak, but no link with this outbreak could be established. One adult who tasted the suspected meal did not become ill. The staff member who had prepared the meal was ill the night prior to the outbreak (vomiting). This person still had abdominal pains the day the outbreak occurred but did not consume the suspected meal implicated in this outbreak.
The suspected meal consisted of rice, which was mixed with cucumber and chicory. The rice was cooked one day before the outbreak occurred, rinsed with cold water, and subsequently stored in a bowl in the refrigerator at the kindergarten for 24 h. Cucumber and chicory were prepared on the day of the outbreak and were mixed with the cold rice prepared the previous day. The mixture was portioned and each portion was reheated individually in the microwave just before feeding the children. Until arrival of the food inspectors, the leftovers of the portions were placed in the refrigerator. The temperature in one of the leftovers at the time of sampling was 5.1°C, which is low as shown by the temperature survey data on domestic refrigerators in the European Union (EFSA, 2012). Samples were sent refrigerated to the NRL-FBO, where analyses started within 6 h following the onset of the outbreak, to prevent further growth and cereulide production in the leftovers.
Laboratory investigation of food, environmental, and human samples
Six food leftovers that had been stored in the refrigerator were analyzed at the NRL-FBO for the presence of B. cereus and coagulase-positive S. aureus, in accordance with the observed symptoms. Microbiological, molecular, and LC-MS2 results are presented in Table 1. All four mashed rice–cucumber–chicory leftovers were shown to contain significant counts of emetic B. cereus (above 1.5E+07 colony-forming units/g), while B. cereus was not detected in the uncooked rice and fresh chicory. B. cereus was also detected in drained liquid from the soaked soiled cloth. Human samples from four cases (three stools and one vomit) were also analyzed: two of the three stool samples as well as the vomit showed the presence of emetic B. cereus. None of the other tested pathogens were detected in any of the analyzed samples.
CFU, colony-forming units; N/A, not assessed; LOQ, limit of quantification (1 ng/g); PCR, polymerase chain reaction; LC-MS2, liquid chromatography tandem mass spectrometry.
The cereulide contents in the four food samples positive for B. cereus, the drained liquid of the soaked soiled cloth, and in the vomit sample collected from a child were quantified by LC-MS2 (Table 1). The detected levels of cereulide varied between 3.1 and 4.2 μg/g in the food samples, whereas only trace levels (below the quantification limit of 1 ng/g) were detected in the liquid from the cloth. The cereulide content in the vomit was 0.350 μg/g.
Discussion
This outbreak highlights the importance of respecting good hygienic practices and strict storage conditions. An ill staff member had been involved in the preparation of the implicated food and may have contaminated the food. Unfortunately, no fecal sample from the staff member could be analyzed and thus her role could not be confirmed. It is recommended to withdraw any ill person from food preparation activities, as stated by the Codex Alimentarius (Codex Alimentarius, 2013).
Moreover, the temperature in the refrigerator where the food was stored had never been registered by the director of the kindergarten. The refrigerator at the kindergarten did not meet the Belgian recommendation on refrigeration temperatures, which are recommended to be between 0 and 4°C, although it was in accordance (<7°C) with the Belgian and European Legislation (FASFC, 2013; RD, 2005). Storage of foods below 10°C prevents the growth of strains that produce emetic toxin, while refrigeration below 4°C is necessary to prevent growth of all types of B. cereus, including psychrotrophic strains (EFSA Opinion, 2005). In this case, high levels of B. cereus were present and cereulide was detected in the leftovers. It is known that cooked rice supports cereulide production at temperatures from 15°C to 37°C (Finlay et al., 2002). Therefore, inappropriate and slow cooling probably allowed the development of B. cereus and subsequent cereulide production in the rice. After cooking, the temperature of the food should be allowed to drop to 10°C as quickly as possible and should ideally reach 4°C to avoid any growth of B. cereus (EFSA Opinion, 2005). Portioning the rice for storage would also have helped to achieve a more rapid drop in temperature below 10°C in the center of the bowl in which the rice was stored.
Subsequent to this outbreak, a plan for cleaning was established at the kindergarten, and measures for temperature control of the refrigerators were taken.
To our knowledge, only a few studies have focused on the actual levels of cereulide implicated in food intoxications, probably due to the previous lack of appropriate quantification methods. Some cytotoxic assays have been used to evaluate the toxin concentrations in food associated with foodborne intoxications; for example, previously reported levels of cereulide were given as a range from 0.01 to 1.28 μg cereulide/g of food (Agata et al., 2002 ) as tested by Hep-2 cell assay. Levels of about 13 μg of cereulide/g of fried rice, as detected by LC-MS2, resulted in multiple ill persons and the hospitalization of an elderly person (Delbrassinne et al., 2012b). In a Belgian lethal case, the level of cereulide in pasta leftovers was 15 μg/g by LC-MS2 (Naranjo et al., 2011). Using the same method, the detected levels in this outbreak were lower but still critical for toddlers, whereas one of the adults who just tasted the contaminated meal did not become ill, presumably because of the lower ingested dose of cereulide per body weight. In this outbreak, the children also started vomiting rapidly after ingestion of the rice, as compared to the reported average of symptom appearance for cereulide intoxication, which is within 1–5 h (Kramer and Gilbert, 1989). Although toxic levels of cereulide have been estimated at 8 and 10 μg/kg body weight using mice and monkeys, respectively, the emesis-inducing dose in human remains unknown (Agata et al., 1994; Shinagawa et al., 1995). Since B. cereus counts and cereulide production may increase rapidly in food over time, it would be important to quantify toxin levels in food leftovers as soon as possible, in an attempt to estimate the intoxicative levels of cereulide for various population groups and to set food safety criteria.
Rapid toxin detection using analytical methods such as LC-MS2 strengthens the evidence of the implicated causative agent even if no vegetative cells can be detected. While rapid detection of toxin in food samples is used in foodborne outbreak investigation, toxin detection in clinical samples might help doctors in their choice of treatment, thus avoiding unnecessary complications such as liver transplantations (Pósfay-Barbe et al., 2008). Indeed, treatment with antibiotics would not cure people intoxicated by preformed toxins, even if presenting the same symptoms as those observed in infections with gastrointestinal pathogens. Moreover, as cereulide causes mitochondrial damage, a more serious outcome such as liver failure can be observed (Dierick et al., 2005; Pósfay-Barbe et al., 2008; Shiota et al., 2010). Unfortunately, no rapid commercial detection methods for cereulide toxin in food and human samples currently exist.
Conclusions
Few investigations of outbreaks caused by B. cereus with supporting laboratory information for toxin levels in food have been described. With use of an available test, the responsible etiological agent for this outbreak was identified and it was shown that moderate levels of toxin could result in illness in vulnerable population groups (in this case, children). It would be recommended to define the toxic doses of cereulide, especially for vulnerable population groups, as the appearance and severity of clinical symptoms might depend on the ingested amount of the toxin and on the overall health status of the consumer. Given that hazardous levels of cereulide can be reached quickly as demonstrated in this investigation, it is important to remain vigilant during food preparation and food storage to prevent illness and outbreaks caused by B. cereus.
Footnotes
Acknowledgments
We acknowledge the technical staff members for all performed laboratory investigations. We thank the Inspectors of the Federal Agency for Safety of the Food Chain and the Health Inspection for the collection of samples and information gathered thanks to their respective inquiries.
Disclosure Statement
No competing financial interests exist.