Sudden Death of a Young Adult Associated with Bacillus cereus Food Poisoning

Abstract

A lethal intoxication case, which occurred in Brussels, Belgium, is described. A 20-year-old man died following the ingestion of pasta contaminated with Bacillus cereus. Emetic strains of B. cereus were isolated, and high levels of cereulide (14.8 μg/g) were found in the spaghetti meal.

CASE REPORT

On 1 October 2008, a 20-year-old man became sick after eating a meal of leftovers of spaghetti with tomato sauce, which had been prepared 5 days before and left in the kitchen at room temperature. After school, he warmed the spaghetti in the microwave oven. Immediately after eating, he left home for his sports activities, but he returned 30 min later because of headache, abdominal pain, and nausea. At his arrival, he vomited profusely for several hours and at midnight had two episodes of watery diarrhea. He did not receive any medication and drank only water. After midnight, he fell asleep. The next morning at 11:00 AM, his parents were worried because he did not get up. When they went to his room, they found him dead.

Legal examination determined the time of death, presumably at 4:00 AM, approximately 10 h after ingestion of the suspected meal. An autopsy could not be performed until 5 days later. Macroscopically, brownish and moderately softened liver and ascites (550 ml of citrine liquid) were found. The heart was macroscopically normal. A total lysis of the pancreas was also found, but it could not be excluded that this finding was due to the autopsy delay.

Microscopic findings were as follows: moderate centrolobular liver necrosis without inflammatory signs and discrete biliary stasis, significant vascular congestion of the lungs, probably due to acute cardiac insufficiency, significant necrosis from all layers of colon mucosa and submucosa alternating with better-preserved zones, and mixed intestinal flora but no evidence of invasive bacterial lesions. Significant lysis of the adrenal glands was also reported. The exact cause of death could not be determined by the autopsy because the interpretation of findings was very difficult due to the autopsy delay.

Five fecal swabs and two feces samples were taken postmortem, and samples were tested for the presence of Bacillus cereus by growth on mannitol egg yolk polymyxin (MYP) agar. B. cereus was found in only two of the five fecal swabs, and the strains isolated were named ISP321 and ISP322. No B. cereus was cultured from the feces samples.

Pasta and tomato sauce samples, the leftovers of the dinner, were also sent for analysis to the National Reference Laboratory for Food-borne Outbreaks (NRLFBO). For enumeration of B. cereus in food samples, the ISO 7932 method (16) was used. Significant B. cereus counts (9.5 × 10⁷ CFU/g) were found in the pasta, while B. cereus was absent in the tomato sauce. The strain isolated from the pasta meal was named ISP303.

PCR assays that detect the presence of toxin genes were applied to DNA from the pasta isolate (ISP303) and the two human isolates (ISP321 and ISP322), and the results are presented in Table 1. The presence of genes encoding nonhemolytic enterotoxin (NHE) (nheA), phospholipase C (plcA), cytotoxin K (cytK), hemolysin II (hlyII), and hemolysin BL (hblA) was investigated (12, 13, 15). All three strains tested positive for the presence of the nheA and plcA genes. The strains were all negative for the other enterotoxin genes (cytK, hlyII, and hblA).

Table 1.

Characterization of the B. cereus isolates^a

Isolate	Source	nheA	plcA	cytK	hlyII	hblA	ces
ISP303	Spaghetti	+	+	−	−	−	+
ISP321	Fecal swab	+	+	−	−	−	+
ISP322	Fecal swab	+	+	−	−	−	+

^aToxin gene presence was assessed by PCR analysis. The “+” and “−” symbols refer to the presence and absence of the gene, respectively.

A specific cereulide-associated PCR test (7) was also used to target the ces genetic determinants, which are necessary for cereulide synthesis. The PCR test to detect ces gave positive results for all three isolates. The actual production of cereulide was also confirmed by the boar sperm assay (4; data not shown). These results suggest that all three B. cereus isolates, ISP303, ISP321, and ISP322, are emetic strains. The concentration of the cereulide toxin in the spaghetti was determined by liquid chromatography-tandem mass spectrometry (LC-MS²) (5; L. Delbrassinne, M. Andjelkovic, A. Rajkovic, P. Dubois, E. N'Guessan, J. Mahillon, and J. Van Loco, unpublished results) to be 14.8 μg/g of pasta. The amount of cereulide determined in the spaghetti by the present study was almost 10 times higher than the cereulide concentration of 1.6 μg/g previously found in a contaminated pasta dish (18) and seems consequently to be within the upper range of cereulide amounts reported in foods implicated in outbreaks.

Further characterization of the three isolates was performed using repetitive sequence-based PCR (rep-PCR) and pulsed-field gel electrophoresis (PFGE) (10, 11, 25) of genomic DNA. All three strains displayed the same rep-PCR and PFGE profiles (data not shown).

B. cereus is a well-known food-poisoning organism. It can cause two types of food poisoning, described as the emetic and diarrheal syndromes (12). The emetic type is caused by a heat-stable cereulide toxin that is preformed in food (2). The symptoms are usually mild but can be severe in some cases. Four fatal cases attributed to cereulide have been reported (6, 22, 24, 26). The present results provided evidence for B. cereus food poisoning in a young, healthy man. Clinical data and the rapid onset of symptoms, together with microbiological and molecular study, point to B. cereus as the most likely cause for this fatal outcome.

Emetic strains of B. cereus are known to cause mitochondrial damage. Because the macroscopical methods used in autopsy could not reveal any specific cardiac symptoms, we performed further, specific microscopical staining. Immunostaining for activated caspase 3, a marker for apoptosis (1, 27), was negative, indicating the absence of mitochondrial damage. The structure of cardiac myocytes was also intact, which suggested that the sudden cardiac arrest was caused indirectly, e.g., due to acidosis.

Fulminant hepatic failure is one of the most feared complications caused by emetic B. cereus strains. In this case, the liver damage was mild and could not be the only explanation for the sudden death of a healthy young man. Therefore, and because of the presence of diarrhea, together with the severe colonic necrosis, PCR assays that detect the presence of enterotoxin genes were applied to the samples.

As indicated above, PCR products were found for the nheA and plcA genes but not for cytK, hlyII, and hblA. These observations are rather difficult to interpret, since the actual implication of putative enterotoxins in diarrhea remains a matter of debate. The absence of HBL enterotoxin in B. cereus strains which produce cereulide is common, whereas the emetic strains usually produce NHE (8). The pattern of toxin production for the strains isolated in the present case is presumably classical for strains implicated in emetic poisoning. Nevertheless, some emetic strains producing NHE were found to be very cytotoxic for Caco-2 cells, as previously reported (14). Furthermore, the simultaneous presence of the genetic determinants for cereulide and the genes encoding potential enterotoxins was recently demonstrated for a B. cereus strain (19, 20), which appeared to be particularly virulent. Could this fact have played a role in the necrosis of the colon, even though the cytK PCR was negative, since that this toxin is the one that has been implicated in necrotic enteritis (21)? That is an important question for which no answer can be given at the moment.

Although we cannot incriminate B. cereus as the direct and unique cause of death, the present case illustrates the severity of the emetic and diarrheal syndromes and the importance of adequate refrigeration of prepared food. Because the emetic toxin is preformed in food and is not inactivated by heat treatment (2, 23), it is important to prevent B. cereus growth and its cereulide production during storage. This toxin production is closely linked to temperature (9) and is not strictly correlated with bacterial counts, as recently demonstrated by Delbrassinne et al. (5; unpublished). The cereulide amounts produced by a B. cereus emetic strain inoculated at 10⁶ CFU/g in cooked rice were higher at 23°C than at 30°C, whereas the opposite situation was observed for the cereulide producer counts. These results suggest that cereulide may be more actively produced at ambient temperatures. In this case, the spaghetti had been kept at room temperature for several days: this allowed B. cereus to grow to the previously mentioned very high concentrations and produced the high toxin concentration (14.8 μg/g) found in the pasta and which is likely responsible for the fatal outcome. According to previously published reports, toxin concentrations that induce emesis in humans, as determined by the cytotoxicity test of HEp-2 cells on the basis of several contaminated foods, extended from 0.01 to 1.28 μg cereulide/g of food (3) and from 0.005 to 32 μg cereulide/g of food (17), determined by LC-MS. Although those concentrations did not implicate lethal cases, we do not know what amounts were actually ingested. In the fatal case described by Mahler et al. (22), strains isolated from the pan residue produced NHE (detected by the BCE-VIA assay) and the emetic toxin when grown on cooked rice. This highlighted the virulence of strains capable of this simultaneous production, as reported in this study.

In conclusion, although the autopsy results were not conclusive, probably due to the delay in analysis, the large number of B. cereus cells and the significant cereulide concentration found in the leftovers are the most likely cause of death of the young healthy man. Further investigation of these isolates and this case may provide insights into the virulence mechanisms of emetic B. cereus isolates.