An Outbreak of Enterotoxigenic E. Coli aboard the M/V Caronia, March 24 - April 5, 2002
- Recommendations and Follow-up Activities
On April 2, 2002, Dr. Elaine Cramer of the Vessel Sanitation Program (VSP), National Center for Environmental Health (NCEH), notified the Foodborne and Diarrheal Diseases Branch (FDDB), Division of Bacterial and Mycotic Diseases (DBMD), National Center for Infectious Diseases (NCID) of a possible outbreak of gastrointestinal illness aboard the cruise ship M/V Caronia. In accordance with illness reporting requirements outlined in the Vessel Sanitation Program Operations Manual 2000, the ship’s physician reported to the VSP gastrointestinal illness in 90 (15%) of 621 passengers and 15 (4%) of 395 crew members by the 9th day of a 12 day cruise. The itinerary of the ship was as follows:
March 24 Passenger embarkation, Acapulco, Mexico
March 25 At sea
March 26 Puerto Quetzal, Guatemala
March 27 At sea
March 28 Puntarenas, Costa Rica
March 29 At sea
March 30 At sea through Panama Canal
March 31 San Blas Island, Panama
April 1 San Andres Island, Colombia
April 2 At sea
April 3 Montego Bay, Jamaica
April 4 At sea
April 5 Passenger disembarkation, Fort Lauderdale, FL
On the basis of higher-than-expected (>3% of passengers or crew) and increasing numbers of passengers and crew reporting gastroenteritis, Cunard Line Limited, collaborated with CDC to conduct an epidemiologic, laboratory, and environmental investigation of the outbreak. On April 3, Dr. Tom Chiller (EIS Officer, FDDB, DBMD, NCID) and Commander Jaret Ames (VSP, NCEH) traveled to Montego Bay, Jamaica, to board the ship and initiate an epidemic investigation on day 10 of the cruise.
The objectives of the investigation were to:
1. Identify the etiologic agent(s) associated with the outbreak;
2. Determine the source(s) of spread of illness among passengers and crew;
3. Formulate interventional strategies to control the outbreak and prevent a recurrence.
A case of gastroenteritis was defined as a passenger or crew member with diarrhea (3 or more loose stools in a 24-hour period) or vomiting plus 1 other symptom (e.g. fever, abdominal cramps, headache, muscle aches, or sore throat) with onset occurring between March 24, 2002, day 1 of the cruise, and April 4, 2002, day 11 of the cruise. A secondary case was defined as a passenger or crew member who met the case definition and who developed illness at least 48 hours after a cabin mate. Secondary cases were excluded from further analyses. Well people were defined as passengers or crew reporting no gastrointestinal symptoms on the questionnaires during the first 10 days of the cruise. People who reported a gastrointestinal illness, but who did not meet the case definition, were classified as having indeterminate illness and were excluded from the cohort study.
The gastrointestinal illness (GI) log, from March 22 – April 3, from the ship’s infirmary was reviewed. Cases from the GI log meeting the case definition were used to generate a preliminary epidemic curve to determine the most likely exposure period. Hypotheses as to the cause of illness were generated by reviewing self-administered food history questionnaires completed by ill passengers who were treated in the infirmary. Menus and daily passenger activities were reviewed with the Executive Chef and the Food and Beverage Manager. All on-board restaurants and other areas where food was served were identified. Dates of shore excursions and description of food served ashore were obtained.
A standardized passenger questionnaire, based on ship menus, shore excursions, and a likely period of exposure determined by the epidemic curve, was distributed to the each of the 621 passengers on April 4th by placing it under the cabin doors. All 395 crew also received a modified questionnaire. Interviewees were asked about demographic information, risk factors for illness, and asked to select food items from a list that they may have eaten during specific meals in the exposure period. Questionnaires were returned to the reception desk on board the ship before passengers disembarked.
Data were entered and analyzed using Epi Info version 6.04 software. Descriptive and cohort analyses of disease onset were based on returned questionnaires from ill passengers and crew who met the case definition. Baseline characteristics of passengers and crew were described by frequencies and percentiles for categorical variables and ranges of values for continuous variables. Univariate analysis was performed using relative risks with 95% confidence intervals (CIs) to test for differences in risk between ill and well individuals. A two-tailed p-value of < 0.05 was considered statistically significant. A cohort study was conducted based upon the entire data set. Exposure variables were coded dichotomously; foods were grouped according to specific meals (lunch on day 3) or specific foods (any chicken consumed during the study period). Exposure variables included shore excursions, ship’s water, ice, and food items from meals throughout the selected exposure period.
Prior to arrival of CDC representatives at the ship, the ship’s physician organized collection of 12 stool samples from ill passengers and crew, and stored the specimens in the infirmary refrigerator. After arriving, 8 additional stool samples from ill passengers and crew were collected. Samples from all stools were put into Cary-Blair transport medium and frozen, and the remaining whole stool was refrigerated. Twenty samples of whole stool and corresponding Cary-Blair tubes were transported in a cooler, on ice, to CDC for analysis. Whole stools were analyzed using electron microscopy and reverse transcriptase polymerase chain reaction. Bacterial specimens were examined for pathogens including the following: Salmonella spp., Shigella spp., Vibrio spp., Campylobacter spp., E.coli O157:H7, Yersina entercolitica, Bacillus cereus, Staphylococcus aureus, and enterotoxigenic E. coli. Selected samples were tested for Norwalk-like virus using RT-PCR with a region “B” primer set.
Samples from cooked baby scallops, cooked baby shrimp and scallop salad, and pre-cooked jumbo shrimp from the galley cold pantry refrigerators were examined at the VSP lab for bacterial content by total plate count for total coliforms, E. coli, S. aureus, Salmonella spp., and fecal coliforms.
The investigative team performed an environmental health inspection of the vessel focusing on the period from March 24, 2002 to April 3, 2002. This included inspection of housekeeping, food safety procedures, and testing of potable water systems. Temperature logs were reviewed for all food refrigerator units; the ship records of the free residual chlorine concentration in the potable water tanks and distribution system were also reviewed. The executive chef and food and beverage manager were interviewed about food service. The ship’s Captain, Staff Captain, and Deck Engineer were interviewed about the potable water system and bunkering of water. The medical staffs were interviewed about recent gastrointestinal illness in the crew. Water samples were collected from potable water tanks, distribution points, and ice machines and analyzed for the presence of bacteria using standard methods for membrane filtration and most probable numbers. Tests were performed looking at total plate count and total and fecal coliforms.
Completed questionnaires were returned by 431 (69%) of 621 passengers and by 340 (86%) of 395 crew. Of passenger respondents, 225 (53%) reported having had symptoms that met the case definition of gastroenteritis; 41 (9%) had indeterminate illness and were excluded from further analysis; 23 passengers met the secondary case definition and were excluded from the analysis; 165 (38%) were well. The median age of ill passengers was 66 years (range 13 – 89 years); 56% were female. Well passengers did not differ in age and sex from ill passengers. Illness among passengers and crew was characterized by primarily diarrhea with accompanying abdominal cramps, nausea, and headache (Table 1). Vomiting was reported by 70 (41%) and fever in 41 (25%). For passengers who had recovered, the median duration of illness was 3 days (range 1 – 9 days); 30% (66/221) reported being ill at the time they returned their questionnaires on day 11 or 12 of the cruise. Of passengers who met the primary case definition, 50% of ill passengers reported onset of gastroenteritis within the first five days of the cruise (Figure 1). The peak date of illness onset for passengers was March 28, 2002, which was the 5th day of the cruise.
Potable water and ice were considered possible vehicles of transmission of disease given the large number of ill passengers with onset of illness spread over many days. Higher gastroenteritis attack rates were observed among passengers who had consumed any beverages with ice before they became ill compared with passengers who had not (Table 2); 185 (58%) of 317 passengers consumed beverages with ice before they became ill, compared with 17 (33%) of 51 passengers who had not consumed any beverages with ice before they became ill [relative risk (RR) = 1.75, 95% CI = 1.17 – 2.61]. In addition, dose-response effects were noted among passengers, with increased consumption of beverages with ice associated with increased attack rates of gastroenteritis (chi-square test for trend = 8.31, p <0.005). There were no differences between gastroenteritis attack rates when comparing passengers who had or had not consumed any unbottled (ship’s potable) water during the cruise (RR = 1.10, CI = 0.73 – 1.67). There were only 2 passengers who drank only bottled water during the study period. There were no significant differences in gastroenteritis attack rates among passengers who went on shore excursions compared with those who did not (Table 2). Shore excursions were not associated with passenger illness. Similarly, eating meals on specific days, eating specific meals, or eating specific food items were not associated with passenger illness.
Gastroenteritis attack rates were also higher among crew members who had consumed any beverages with ice compared with crew members who had not; 18 (19%) of 196 crew members who consumed any beverages with ice became ill, compared with 6 (4%) of 141 crew members who had not consumed any beverages with ice (RR = 2.13, CI = 0.87 – 5.23).
Whole stool samples were collected from 14 passengers and 6 crew; 2 swabs for bacterial analysis were obtained from each of the whole stool specimens. Three specimens exhibited no growth. Enterotoxigenic E. coli (ETEC) was isolated from eight passengers and one crew member (Table 3). Multiple ETEC serotypes were isolated: O27:H7 (5), O148:H28 (2), O79: Hund (1), O25:NM (1), O6:H16 (1). All of the serotypes had the gene sequence for the heat-stable toxin (ST) and two had gene sequences for both the heat-stable toxin and the heat-labile toxin (LT/ST). Shigella sonnei was also isolated from one passenger. All 20 cultures were negative for Salmonella, Yersinia, Campylobacter, E. coli O157:H7, and Vibrio species.
Ten stool samples, five that were positive for ETEC, from seven passengers and three crew, tested negative for Norwalk-like virus.
Record review and food service inspection with the Food and Beverage Manager and Executive Chef revealed hot/cold food storage temperatures and preparation practices in compliance with VSP sanitation requirements. Food handlers wore disposable gloves and aprons as required.
During the cruise, water was bunkered (potable water taken on in port) in Acapulco (on 3/24), Guatemala (3/26), Costa Rica (3/28), Panama (3/30), and Jamaica (4/3). On the M/V Caronia, bunkered water is routinely tested for the presence of total coliforms and E. coli using Colilert test kits before and immediately after chlorine injection. The water bunkered in Guatemala tested positive for total coliforms and E. coli pre-chlorination and tested positive for total coliforms and negative for E. coli post- chlorination. No further treatment was administered prior to distribution. From potable water chlorine analyzer records, potable water chlorine levels were observed to be appropriate (between 0.5 – 2.0 ppm) at the distribution system level during the entire cruise, never dropping below the 0.2 ppm minimum requirement. Chlorine was added to the bunkered water at the required 2 mg/L. This was increased to 3 mg/L after the positive test from the water bunkered in Guatemala.
Ice machines delivered ice in a manner where ice handling was not considered to contribute to contamination. There was however, one ice machine on the lido deck buffet where the trigger contacted the lip contact surface for the cups. A sign posted advised passengers not to refill used cups, but passengers were observed re-using cups. Activated charcoal filters were in place on all ice machines to remove chlorine from water before it was frozen into ice. Samples of water (n = 9) and ice (n = 4) taken from the ship on April 5, 2002 and tested by VSP tested negative for the presence of coliforms.
A large outbreak of gastroenteritis caused by ETEC occurred among passengers and crew aboard the M/V Caronia during a cruise from Acapulco, Mexico on March 24, 2002 to Fort Lauderdale, Florida on April 5, 2002. Clinical characteristics of illness were consistent with ETEC, which was identified among 9 of 20 stool samples collected from ill passengers and crew. The peak of the outbreak occurred on the fifth day of the cruise. Given that the typical incubation period for ETEC infection is between 21-48 hours, the initial exposure most likely occurred in the first several days after embarkation. Questionnaire data revealed a statistical association between consumption of beverages with ice and illness among passengers and a dose response related to the number of iced beverages consumed. A similar association was seen among the crew members, although it was not statistically significant probably because of the small number of cases among the crew.
The shape of the epidemic curve (Figure 1) is most likely due to several factors. In this case, contaminated water contained multiple pathogens resulting in several distinct incubation periods which are represented in the shape of the epidemic curve. Furthermore, the ship offered free health visits for anyone with diarrhea beginning on March 31 which explains a rise in case numbers identified.
The most likely source of contamination of the ice was the water used in the ice makers. Based on the onset dates, this could have been water bunkered in Acapulco or Guatemala. Findings from the environmental investigation demonstrated that the potable water bunkered in Guatemala was contaminated with coliforms before and after chlorination. This suggests that water was contaminated with enough fecal material to survive standard chlorination and allowed ETEC to be introduced into the ship’s water system. Furthermore, the multiple serotypes of ETEC found in stool samples from ill passengers and crew support the finding that there was fecal contamination of the bunkered water.
Epidemiologically, ice was associated with illness, but not the ship’s water. One hypothesis accounting for this is that ETEC reached the ice machines through the water in conditions of inadequate chlorination. There, the filters removed chlorine prior to freezing. ETEC could then have remained viable in the non-chlorinated ice. The same water may have had longer chlorine contact time in the potable water system allowing for better disinfection. The fact that ice and water samples taken by VSP were negative for the presence of coliforms at the time of the investigation is not surprising given the additional chlorine added to the water system upon identification of the outbreak. It is unlikely that ice was contaminated by ice handling, given that ice machines used on board the ship had separate dispensers and trigger buttons.
On April 12, 2002, VSP made the following recommendations to Cunard Line Limited for immediate implementation:
I. General recommendations for the control of gastroenteritis aboard the next cruise:
1. Continue to submit daily reports of all diarrheal illness over the next several cruises to monitor for cases of gastroenteritis.
2. Encourage both passengers and crew with diarrhea to immediately report illness to the infirmary. Obtain and label stool samples from all passengers and crew who have diarrhea and refrigerate immediately.
3. Reinforce the importance of good hand washing practices among all crew.
4. Exclude food handlers with vomiting or diarrhea or a history of vomiting or diarrhea until they are symptom-free for 72 hours before returning to duty. Cabin mates of sick crew should be confined to cabins. If they experience no illness after 48 hours they can return to work.
5. Establish procedures for handling soiled crew cabin laundry separately, as is the case for soiled passenger laundry. A separate machine should be used with high heat and chlorine added.
II. Specific recommendations for management of potable water on the M/V Caronia:
1. Bunker water only from reliable sources of well documented quality. Specifically avoid bunkering in Guatemala on future cruises.
2. Increase free chlorine residual to 3 ppm or greater when bunkering potable water.
3. Maintain free chlorine residual at 1 ppm at the far point of the distribution system during the next cruise.
4. Monitor and record the tanks filled during bunkering and record a quantity in tons for each tank.
5. Conduct microbiological tests on all potable water being bunkered. Water that is positive should be stored and the free chlorine levels should be raised to 5 ppm and the tanks retested after chlorination. If water is still positive after initial chlorination, water should be held for 24 hours and retested. No water should be distributed that is still positive. If possible, any water testing positive should be dumped.
6. Empty all ice machines, clean, sanitize, and make new ice.
7. Eliminate passenger self-service in the lido buffet by posting staff to serve passengers food from the serving line and drinks from the beverage station. Repair or replace the ice maker/dispenser so the trigger is separate.
8. Thoroughly disinfect all passenger garbage receptacles and sanitize ice buckets. Use disposable plastic liners in ice buckets and trashcans in all cabins.
Cunard Line Limited, staff and crew of the M/V Caronia cooperated fully with this investigation. They facilitated distribution and collection of passenger and crew surveys, collection and storage of stool samples, provided copies of updated environmental and medical records, and implemented recommendations listed above.
Tom M. Chiller MD,MPH
Epidemic intelligence Service Officer
Foodborne and Diarrheal Diseases Branch
Division of Bacterial and Mycotic Diseases
National Center for Infectious Diseases
Elaine H. Cramer, MD,MPH
Vessel Sanitation Program
Division of Emergency and Environmental Health Services
National Center for Environmental Health
Table 1 Characteristics of clinical illness (from 225 ill passengers and 24 ill crew members), MV Caronia cruise, March 24- April 5, 2002.
Table 2 Gastroenteritis attack rates among passengers associated with exposure to selected risk factors, MV Caronia cruise, March 25 – April 5, 2002.
Table 3 Stool specimen ( n = 10) results from passengers and crew members with gastroenteritis aboard the MV Caronia cruise, March 24 – April 5, 2002.
Figure 1 - Date of illness onset among passengers aboard the MV Caronia cruise, March 24 – April 5, 2002