The recent outbreak took place on a cruise carrying 435 pensioners; several were affected. The total number of sufferers, according to a questionnaire returned by only 339 of the cruise goers, was 219 (64.6%). This indicates that point source of the outbreak must have come from the cruise itself. This called for an investigation and an epidemiological test to be made in order to get answers. Known food samples of the original meals that were served on the cruise were diluted and inoculated onto (PEMBA) spread plates for incubation.
In addition to the suspected food causatives, body substance samples which would have come into direct contact with the suspicious food samples were thought to have had an influence prior to the outbreak occurring and were therefore also plated out. Different media were used to plate out these samples;
Faecal specimen from victim: PEMBA
Xylose Lysine Desoxycholate agar (XLD)
Swab from cook's nose: Blood Agar (BA)
Mannitol Salt Agar (MSA)
After at least 24 hours (48 hours for PEMBA) of incubation, initial identification and confirmatory tests as well as colony forming unit (cfu) counts were performed on colonies from each medium. Further and more definitive tests were run on the second day depending on results (aka suspicious colonies); O/F medium of Hugh and Leifson, DNAse plates, and latex coagulase tests were set up for further incubation.
The results for all food and faecal sample spread plate identification tests are shown in Table 1 and the results from the cook's nose plate identification test are shown in Table 2.
Although numerous tests were performed on each of the isolates as well as the cook's nose sample, only two organism identification was made. It was determined that Staphylococcus aureus was present in the sample of chicken isolate, and Corynebacterium in the lamb isolate.
In addition to the results displayed in the tables below, Table 3 shows supplementary statistical interpretations of relative associations happening between the food specimens and illness that were performed using chi-squared values and odds ratios for each food sample.
Table 1:Results of known food and faecal sample spread plate identification tests.
Chicken
Ham
Roast lamb
Rice Pudding
PEMBA count cfu/gm
~ 800
≥ 300
Scant
TNTC
Replicates: 1.3x105; 1.7x105; 1.2x106; 4.1x106
PEMBA col. appearance
Tiny, white
Tiny, yellow/white
1mm white
Small yellow
Large 4-5mm, blue, matt, irreg. margin, halo of ppte. Some proteolytic clearing.
PEMBA isolate tested:
Gram
GPC clusters
GPC
Small GPR
GPC
Large GPR in chains
Spore stain
oval central spore +
Lipid
lipid granule pos.
Catalase
cat. Positive
cat. Negative
cat. Positive
cat. Negative
Cat. Positive
motility
motile
Food isolate - PEMBA subculture
pure, large blue, matt
Glucose VP
VP pos.
Glucose anO2
anO2glu: grew/acid +
Indole
indole neg.
Gelatin
gelatin pos.
Urea
Urea neg.
Growth 45o
Growth pos
Growth 65o
No growth
Starch hydrolysis
Stratch pos.
Casein proteolysis
Casein neg
Nitrate test
Nitrate pos. no gas
(BHIA purity)
pure
API identification
? staphylococci
? diphtheroid GPR
?Corynebacterium
Faecal sample from person eating that food
PEMBA
Small, white colonies. GPC; cat. neg
Tiny yellow colonies. GPC cat.neg
Tiny yellow colonies GPC; cat. neg
Small yellow col.
Large, yellow/blue matt cols. (s/c to PEMBA yielded large, blue, matt col.)
XLD
Med. Yellow
Orange/red (H2SNEG) col.;s/c to MAC gave LF
Med. Yellow cols.
Med. Yellow cols
Med. Yellow
Orange/red (H2SNEG) col.l s/c to MAC gave LF
Table 2: Results of body substance plate identification tests (the cook's nose).
Cook's nose swab
HBA
Weak, β-haemolytic, small, round, flat, white colonies
MSA
Small white-ish colonies, pink agar throughout.
BA-isolate:
Gram
GPC, clusters
Catalase
Positive
Latex coagulase
… (no visible clumping)
O/F
… (yellow throughout tube)
DNAse
… (agar green all around colony - no halo of clearing like pos. control)
ND: not determined TNTC: too numerous to count +: positive neg.: negative s/c: subculture
MAC: MacConkey agar LF: lactose fermenter O/F: oxidation/fermentation test (Hugh & Leifson)
DNAse: deoxyribonuclease cat.: Catalase
API Staph
ND
By implementing the use of chi-squared values and odds ratios, the probability that a significant association between the particular isolate and the illness is what's being assessed. (Table 3)
Food Sample
Ill
Not Ill
Chi-Squared (X2) value
Degree of Freedom (DF)
P-value
Odds Ratio
Confidence Interval (95%)
Chicken
Eat
127
120
10.009
1
0.001
<0.05
0.44
0.26 to 0.74
Not Eat
65
27
Ham
Eat
115
87
1.637
1
0.201
>0.05
0.62
0.39 to 0.98
Not Eat
77
43
Rice pudding
Eat
190
86
10.433
1
0.001
<0.05
5.46
1.14 to 26.1
Not Eat
2
8
Roast lamb
Eat
169
68
1.277
1
0.258
>0.05
0.77
0.38 to 1.56
Not Eat
23
14
Table 3: Statistical interpretations of relative associations between the food specimens and illness.
For an outbreak as such, a null hypothesis (H0) would be produced suggesting that illness is independent of risk factor (that is, the isolate), as well as an alternative hypothesis (H1) suggesting that that illness is not independent of risk factor - in turn suggesting a relationship between the two. In the table created for this particular outbreak, where chi-squared was calculated and tabulated showing differences between whether the food eaten or not had an impact on illness or not, the results suggest several theories. If the reported p-value is less that 0.05, H0 is rejected and H1 accepted; moreover if p < 0.05, the statistical interpretation of the relative association between the food specimen and illness is significant. This significance is then further tested by means of an odds ratio. When values less than 1 are shown, it means there is no real risk (or no increase in risk) when choosing to eat the particular food in the group tested. However, if the value of odds ratio is larger than 1, then the risk of illness is respectively increased through exposure to that food. Table 3 shows that the rice pudding (odds ratio: 5.46) was the one particular type of food served on the cruise which has a significantly higher risk factor (5-fold) of illness than the other specimens.
Discussion:
In order to associate fully a particular isolate with the illness, certain tests must be done in the laboratory to differentiate it from the other isolate and come out with an explanation as to what the causative agent for the outbreak was. The way growth on plates appears, particular tests for positive or negative results as well as confirmatory tests are all key to investigate the organism that is epidemiologically significant in this outbreak.
From the results of the initial stages of the experiment, it was clear that the specimen of the rice pudding had some form of unusual growth in it (Table 1) not displayed by the other food types; with a colony forming unit (cfu) count significantly exceeding that of the other specimens - over 1.3x105. It wasn't just the large number of colonies; it was apparent that the colonies were all of the same morphology without any real secondary colonies. They were large, blue, matt and irregular in shape with a margin and halo of precipitate around each colony.
Based on observation by the naked eye alone, it was determined that the growth on the PEMBA by the rice pudding sample might have been Bacillus cereus due to the size, colour and shape of the colonies. "After 18-24 hours of incubation at 30+/-1°C in aerobic conditions, Bacillus cereus shows crenated, colonies about 5 mm in diameter, turquoise blue in colour, surrounded by a distinct opaque zone of egg yolk precipitation of the same colour as the colonies." (ICMSF, 1978) The action of B.Cereus is known to cause a clearing of media that contains egg yolk emulsions; this results in a characteristic halo around the colony. This is exactly what was seen on the PEMBA plates in the lab suggesting that the growth might indeed be Bacillus cereus.
However, this alone is not enough for determining causation of an outbreak and therefore further tests were conducted on the rice pudding isolates. Each isolate was subjected to a Gram Stain exam, a catalase test, lipid and spore stain tests as well as motility testing. Unfortunately, colony results for 3 of 4 specimens tested happened to be inconclusive.
There was growth from the chicken and, when tested under a microscope, showed gram positive cocci in clusters suggesting the possible presence of staphylococci in the chicken sample. Statistics tests also showed that the p value showed a significant relationship between eating the chicken and getting ill. This value was the same for the rice pudding which also showed large gram positive rods when a microscopy was conducted on the isolate; this adds more affirmation that the colonies were indeed Bacillus, aka "significant". (Neal, 1997)
Chi-Squared (X2) value
P-value
Odds Ratio
Chicken
10.009
<0.05
0.44
Rice pudding
10.433
<0.05
5.46
It is worth noting that a p value less than 0.05 would result in the rejection of the null hypothesis - therefore showing significance: (About.com)
The ham isolate showed a smaller number of colonies, suggesting that there was no real risk factor associated with the ham - this is also shown in Table 3, where the p value proved insignificant. From the lamb isolate, different student came up with different results, however the statistics given at the start of the outbreak, when tabulated with chi-squared and odds ratio, show a not-so-significant relationship between eating the lamb and getting ill (p value >0.05) - and the different results by different students suggests an inconsistency in results - therefore further tests were also not needed.
As mentioned, the determination of causative agent for illness (based on colony presentations alone on the PEMBA plates with rice pudding isolates) to be B.Cereus was not enough to be certain and so further tests needed to be done. Sub-cultures of the PEMBA isolates were tested are results for all these tests are shown in Table 1.
Table 1 shows VP with positive results; indicating the production of acetylmethylcarbinol. This result is characteristic for B. cereus. Another confirmation of B. cereus is when the Phenol red glucose broth changes colour - this happened and indicated that acid had been produced anaerobically from glucose. Gelatin is a protein that B. cereus is able to hydrolyse, thus, nutrient gelatin media which is being inoculated with B. cereus should stay in the form of a liquid when put inside an ice-cold bath. B. cereus has proteases that cause gelatin to hydrolyse and therefore giving a positive result (just as with the results in this experiment). (De Clerck et al., 2004)
The best temperature for the growth of B. cereus is anywhere between 30°C to 50°C (Hocking et al., 1997) - this is shown in the results when the temperature was at 45°C, growth was noted. However at a higher temperature (65°C) there was no growth of the bacteria.
Starchy foods are known to be places where B. cereus grows (Hocking et al., 1997) and according to the results, the sub-culture of the rice pudding specimen displayed a positive result for starch hydrolysis. This is another breakthrough into the discovery of the outbreak causation.
The nitrate test shows a positive result for the rice pudding sub-culture isolate; this is another reason to assume that B. cereus is present. The bacteria can be identified in isolates by testing whether nitrate gets reduced to nitrite to attain a positive result.
In order to determine if an organism is present in the body's natural flora, faecal samples from the patients who are the foods were also provided. However, error was expected as most people ate more than one type of food. Though, for the experiment, it was assumed that the results were all separate. The only suspicious colonies were the rice pudding isolates on PEMBA, as well as the chicken and rice pudding isolates on the Xylose lysine deoxycholate (XLD) agar. The growth for both the ham and lamb isolates, were insignificant as Staphylococcus aureus is not always pathogenic. It is worth mentioning that although S. aureus happens to be a known enteric pathogen, it is also a major part of the body's natural outer flora. Finding it present in specimens does not necessarily indicate pathogeneses or illness, though this may be the case depending on the location of the organism. This is the main reason as to why these tests were run on the affected patients' faeces.
The PEMBA plate, which had the rice pudding isolate on it showed, visible evidence of B.Cereus growth due to the size, colour and shape of the colonies. The XLD plates for chicken and rice pudding showed orange/red growth which might suggest Shigella, however when a sub-culture was tested against a MacConkey agar (MAC), the results displayed a pink colony (LF) which suggests that Shigella is not present; hence this assumption was quickly ruled out. Other suspected bacteria were also ruled out such as Salmonella.
The gram positive cocci on the Baird Parker isolate from the cook's nose could suggest that Staphylococcus aureus was present; however this isn't necessarily a causative agent for the outbreak as following tests proved non-determinable.
"According to Koch's postulates:
(i) The microorganism must be present in every case of the disease but absent from healthy organisms.
(ii) The suspected microorganisms must be isolated and grown in a pure culture.
(iii) The same disease must result when the isolated microorganism is inoculated into a healthy host.
(iv) The same microorganism must be isolated again from the diseased host." (Prescott et al, 2008)
Ultimately, due to the nature of the study conducted being one that is retrospective; students at UTS were unable to fully test and provide solid proof that the organism was, in fact, the causative agent according to Koch's postulates. However, as B.Cereus was detected as present in the food that proved most significant in relation to the number of people who fell ill (rice pudding), and it was also found present in the faeces of the patients who are the rice pudding samples as well. Due to these observations and results, the students were able to come up with a causative agent. This is especially true when combined with the statistical results in Table 3.
At this stage it is safe to say that Bacillus Cereus was the causative bacteria in this outbreak, however the question remains; 'how did it get there?' According to the UTS lab manual, the outbreak took place onboard a harbor cruise; however the onset of symptoms began within hours of the first person ingesting the food. The origin of the bacteria is undetermined, but could have come from the rice which might have not been cooked properly. (Turnbull, 1996) B.Cereus causes nausea, vomiting and diarrhea - all of which were symptoms described by the patrons onboard the cruise. Low temperatures of cooking could result in the survival of the bacteria's spores. Leaving the rice pudding in a fridge only helps the endopores to germinate and become infective (10°C -50°C). (Roberts et al., 1996) This might have been the case as the food would have had to be transported from land onto the cruise, and might have been cooked and kept chilled prior.
The onset of illness began within hours of eating the food. There was a small number of cases within the first 3 hours, however most of the incubation period happened after 8-12 hours of ingestion. The first few cases might have been related to the fact that vomiting is often caused by rice that is not cooked well enough (right time and temperature) to kill any spores, then kept cool. This initial onset might not have neceserily been caused by B.Cereus, but by the Staphylococcus aureus which was found in the ham and lamb - however there isn't enough evidence to prove this. (Roberts et al., 1996)
The diarrhea that followed after several hours would have definitely been due to the B.Cereus present in the rice pudding. The regular incubation period for this bacteria is 8-17 hours and that is exactly when most of the symptoms were recorded to have taken place.
It can now be noted that the timeline of events started off with the rice being improperly cooked, then kept in a fridge to cool down - this caused the bacteria to grow. Once the rice pudding was on board the ship, and people started eating it, the illness started to spread. The equipment used to scoop the pudding might have contaminated other foods (chicken, ham and lamb) and so this might have been why most of the people on board got ill. The cook's nose was held suspicious because it might have come from him; however this is unclear as results didn't show much.
Ways of prevention would be to make sure the rice is cooked properly and not kept at such low temeratures. Also , it is important to keep a clean environment around the kitchen and cooking area.
