While there was no smoking gun, there were a number of issues that may have led to the outgrowth of this sporeforming pathogen. As you know, spores of Clostridium perfringens will survive cooking, and if that product is temperature abused, the spores will germinate and the organism will multiply very rapidly.
The issues revolved around temperature measurements during holding of the food. The recommendations in the report provide good guidance for operations to prevent such outbreaks.
From the report:
Recommendations for event A and all event organizers and Caterer A and all foodservice facilities:1.) Ensure that internal food temperatures are measured at the conclusion of cooking and during the hot holding process.Maryland Department of Health
a. Temperatures should be taken while the food remains inside the hot holding cabinets at one hour intervals and from multiple locations of the food trays on different shelves.
b. Food handlers should record the range of temperatures (versus a single temperature) as observed on log sheets. Food must maintain 135°F at all times after cooking and prior to service.
c. Obtain representative (multiple sites, mix of locations on tray, such as center, corners, edges) temperature measurements of all food trays before serving time.
2.) Report immediately to management or the person in charge when any food temperatures are below the required holding temperatures.
3.) Corrective action, as specified in the facility’s approved HACCP plan, must be taken when food measures less than the 135°F critical limit.
4.) Maintain detailed temperature logs.
a. Retain detailed internal temperature logs of any cold and hot held food every hour for all locations and all serving lines;
b. Log both internal and external temperature readings for all refrigeration units every 2‐4 hours to ensure that potentially hazardous foods do not exceed regulated time and temperature requirements.
http://dhmh.maryland.gov/docs/Outbreak%202014-119%20FINAL_with%20Attachments_v3.pdf
SUMMARY REPORT
OUTBREAK 2014-119
September 2014
Office of Infectious Disease Epidemiology and Outbreak Response
Prevention and Health Promotion Administration
Maryland Department of Health and Mental Hygiene
INTRODUCTION
On April 11, 2014, the Baltimore City 311 system received 3 reports of illness from attendees of Conference A. A 4th report was received on April 15. All of the reports were from conference attendees who also worked in the same building at another work location. The reporters stated that they, and several coworkers who also attended Conference A, became ill with diarrhea between April 8 and April 10. The attendees suspected that lunch served on April 9 was the source of the illnesses. All 4 reports were assigned in the 311 system to Baltimore City Health Department’s (BCHD), Bureau of Environmental Health, Environmental Inspection Services (EIS) Food Control Section. On April 16, BCHD, EIS identified that these reports were related and informed BCHD’s Office of Acute Communicable Diseases (ACD). An outbreak investigation was initiated on April 16 by BCHD. BCHD notified the Maryland Department of Health and Mental Hygiene (DHMH) Division of Outbreak Investigation on April 16. Subsequently, the response proceeded as a joint state‐local outbreak investigation.
BACKGROUND
Conference A was held at Convention Center A in Baltimore, Maryland. The main conference was held from Tuesday, April 8 through Thursday, April 10. Two smaller training sessions were held on Monday, April 7. Approximately 1300 people attended, exhibited at, or spoke at the conference. Attendees were from 42 states, Canada, Mauritius, and Costa Rica.
Caterer A, the primary caterer for Convention Center A, supplied food for the conference. Food was also available for purchase at vendors and concession areas in the convention center. Caterer A provided food for continental breakfasts on April 8, 9, and 10, lunch on April 9 and 10, afternoon break on April 8, 9, and 10, and for evening receptions on April 8 and 9. All food was served buffet style.
METHODS
Epidemiological findings [See report]
Environmental findings:
Food was not being prepared in the kitchen during the initial visit on April 16 by two BCHD Environmental Health Specialists. The observed temperature of a walk‐in cooler was within the acceptable range. No food prepared for the conference remained. Menus for the conference, temperature logs, and recipes and procedures for food preparation were obtained.
During the same visit on April 16, Caterer A reported that during the event, about 12 employees worked in the kitchen and 40 served food. Of these, Caterer A estimated that about 20 were temporary wait staff. Two other events were held at Convention Center A that same week and were also catered by Caterer A, but neither the facility nor Caterer A received any complaints of illness from either of these groups. DHMH called the contact person that was supplied to the convention center for one of the groups to ask if attendees of that event had become ill. Caterer A also reported that two employees had been ill around the time of the outbreak. One employee who washes dishes started feeling sick the morning of April 10 and vomited in a restroom at work that afternoon; the employee was sent home.
Another employee, who did not work in the kitchen, became ill with diarrhea and an upset stomach on April 11, with duration of symptoms less than one day. According to Caterer A management, neither of the ill employees ate food from work. Both of the ill employees submitted stool specimens for testing.
Also during the April 16 visit, Caterer A reported hearing rumors that 20 people associated with Conference A were sick and that someone working at the registration desk for the conference had been sick and could have contaminated the attendees’ badges. When additional questions were asked about this at a later date, Caterer A reported that it was the event organizer who told them of the illnesses on April 10. The event organizer and caterer did not report the illnesses to BCHD or DHMH.
Through subsequent calls, emails, and visits with Caterer A, information about the preparation and holding of the chicken Marsala served on April 9 for lunch was obtained. The Marsala sauce was prepared the morning of April 9 and used only for the April 9 lunch. Kitchen staff might also have consumed the chicken Marsala, but the other two groups with events at Convention Center A that day had a different menu and would not have eaten it. Leftovers would have been discarded. However, the ingredients used to make the dish were likely used for other dishes served to this group and for dishes served to other groups. Caterer A reported that precooked frozen chicken breasts were used for the chicken Marsala. The chicken breasts were placed on sheet pans and thawed in a walk in cooler on April 8. They were cooked the morning of April 9 and transferred to 2‐inch pans after cooking. The sauce was prepared using wine, pre‐sliced fresh mushrooms, 16 lb. buckets of frozen demi‐glace that had been placed in a walk‐in cooler 24 hours prior to thaw, and peeled fresh garlic from 5 lb. jars that was chopped in the kitchen prior to use. The mushrooms and garlic were added first to a steam jacket kettle, followed by the wine and demi‐glace. The sauce was brought to a boil and then simmered for 30‐40 minutes. One hundred and fifty gallons of sauce were prepared at one time. The sauce was drained from the steam jacket kettle into pitchers and poured directly over the pans of cooked chicken breasts.
Plastic wrap was placed over the pans. The pans were loaded into hot holding cabinets with Sternos on the bottom shelf approximately 1 hour and 20 minutes prior to service. Temperatures were recorded at that time and 2 hours later. The hot holding cabinet was plugged in while in the kitchen, unplugged during transport, and plugged in at the location of service. Transport to the location of service occurred about 50‐60 minutes prior to the opening of the buffet. Fifteen to 20 minutes before the buffet lines opened, the pans were loaded onto pre‐warmed serving dishes with the lids closed. The plastic wrap was removed when the buffet line opened. The above process was reported by Caterer A management and not directly observed on the day of service by DHMH or BCHD. Time intervals were reported, not recorded, by Caterer A.
Three temperature logs for the April 9 lunch service were available from Caterer A. Temperatures were recorded for 2 time points. All three logs were similar and indicated a temperature of 167°F at 10:15 am and a temperature of 151°F or 152°F at noon for the chicken Marsala. Temperatures were also recorded for vegetable lasagna and roasted vegetables. Per Caterer A, the buffet was scheduled to be open from 11:45 am to 1:15 pm.
Laboratory findings:
Stool specimens from 22 ill individuals were tested, including 1 from an individual who was excluded from the exposure analysis because their onset of illness was April 17, 2 from ill employees of Caterer A, and 19 from case‐patients. The median time between onset of illness and specimen collection was 13 days (range 4‐21 days). Two specimens from case‐patients were positive for norovirus genogroup 2 by real‐time reverse transcriptase polymerase chain reaction procedure (RT‐PCR). The specimen from the person with an onset of illness on April 17 who was excluded from the exposure analysis was positive for norovirus genogroup 1. Testing for Salmonella, Shigella, E. coli O157, Shiga toxins, Campylobacter, rotavirus, sapovirus, astrovirus and Shiga toxins was negative.
Four of 17 stool specimens were positive for CPE toxin by PET‐RPLA. (Diagram 2.) Of those 4 specimens, suspect colonies were observed for 3. Of the 3 with suspect colonies, all were cpa positive but were negative for cpe and cpb. Thirteen of 17 specimens were negative for CPE toxin. Of those 13, suspect colonies were observed for 7. Of the 7 with suspect colonies, 1 was cpa positive, cpe positive, and cpb negative. The other 6 were cpa positive, cpe negative, and cpb negative. No suspect colonies were observed for 6.
Frozen chicken breasts, garlic, and demi‐glace were cultured for B. cereus and C. perfringens. Results of <10/gram were reported for both types of bacteria in all samples, indicating that these bacteria were either absent or possibly present at very low levels, below the limit of detection.
CONCLUSION
In this gastroenteritis outbreak associated with Conference A, the majority of cases appear to have been associated with a point source exposure. Over half of the cases (117 of 216 total cases) had onsets within a 16 hour period (4 pm on April 9 until 8 am on April 10). Conference attendees who ate any food from the lunch served on April 9 were over five times more likely to develop illness than attendees who did not eat lunch on April 9, suggesting that exposure for these cases may have occurred during that meal. Of the food items served during that meal, chicken Marsala was most strongly associated (RR= 3.5 (2.0, 6.1)) with an increased risk of illness and was consumed by a majority of cases.
The frequency of signs and symptoms, duration of illness, and possible incubation period were consistent with outbreaks caused by C. perfringens. In this outbreak, almost all of the cases had diarrhea and only 10% reported vomiting, which is typical of outbreaks caused by C. perfringens. The median duration of illness was 28.5 hours, which was slightly longer than is typical for outbreaks caused by C. perfringens. The median onset of illness for the outbreak, coinciding with the peak of the epidemic curve, was at 12 am on April 10, approximately 12 hours after the April 9 lunch which was epidemiologically implicated. If the majority of cases were exposed to an etiologic agent during that lunch, the median incubation period for that agent would be approximately 12 hours, which is consistent with C. perfringens. These incubation period findings were similar to incubation periods described with other potential etiologic agents, including B. cereus, and such pathogens cannot be ruled out.
In this outbreak, there were other reported gastrointestinal illnesses that did not appear to be related to the majority of cases. Two of 19 specimens from case‐patients tested positive for norovirus; however, norovirus was unlikely to be the etiology for the majority of cases. Given experimental evidence (1) and experience with previous outbreaks it would be expected that if norovirus had caused most of the illnesses, a higher proportion of specimens would have tested positive for norovirus. Therefore the laboratory results taken as a whole are not inconsistent with a predominant etiology of C. perfringens.
It is certainly possible that some transmission of norovirus occurred at the convention and accounted for illness in some of the cases, and some of the cases may be attributable to background levels of norovirus circulating in the community at the time of the outbreak. In support of this, one individual who tested positive for norovirus genogroup 2 was likely infected prior to attending the conference, given the short period between attendance and onset.
If norovirus rather than C. perfringens or B. cereus, had caused illness for the majority of cases, it would be expected that the epidemiological features of the outbreak would have been different. C. perfringens and diarrheal type B. cereus typically cause diarrhea and cramps without vomiting lasting less than 24 hours. (2) (3) A review of U.S. C. perfringens outbreaks reported from 1998‐2010 found 91% of cases had diarrhea and 14% had vomiting, consistent with the signs and symptoms reported by the patients in this outbreak. (4) Incubation periods for diarrheal type B. cereus are typically 6 to 15 hours, (3) and incubation periods for C. perfringens are typically 8 to 12 hours (range 6 to 24 hours). (2) In contrast, outbreaks caused by norovirus usually meet the Kaplan Criteria: more than 50% of cases have vomiting; the average duration of illness is from 12 to 60 hours; the average incubation period is from 24 to 48 hours; and no bacterial agent is found. (5)
Our epidemiologic findings suggest that a single exposure and agent accounted for the majority of the cases associated with this outbreak. The vehicle for infection appears to have been the chicken Marsala dish served at the April 9 lunch. Chicken Marsala could be a suitable medium for C. perfringens and B. cereus, and under certain conditions, growth sufficient to cause illness could occur relatively quickly.
Meat, poultry, and gravy are foods commonly associated with C. perfringens outbreaks (2), and chicken Marsala (chicken breasts served with gravy) has been implicated in previous outbreaks. (6) C. perfringens bacteria are commonly found in the intestines of humans and other animals, on raw meat and poultry, and in the environment; hence, it would not be unusual to find low levels of contamination on food before cooking. (2) The spores can survive cooking and germinate afterwards, so it is also not uncommon to find small numbers of bacteria on just‐cooked food. (3) If food is left at temperatures between 54°F and 140°F after cooking, growth can occur, with the most rapid growth occurring between 109°F and 117°F. (2) With a fast doubling time of less than 10 minutes, counts of C. perfringens can reach high enough levels to cause illness relatively quickly compared to other foodborne pathogens. (3)
When large quantities of live C. perfringens bacteria are ingested, they can produce a toxin that causes diarrhea and cramping in the intestine. (2) If the chicken Marsala was not continuously held above 140°F after cooking, growth of C. perfringens could have occurred.
It is difficult to determine what exactly happened in retrospect, however, at gatherings and events such as this one where large groups of people are served and large quantities of foods such as roasts, gravies, and poultry are often cooked in large batches or prepared ahead of serving, it is possible that proper cooking, cooling, and hot holding intended to decrease the growth of C. perfringens and other bacteria in food may not have been achieved. Although the temperatures recorded on the logs were within the acceptable range, it is not possible to rule out that temperature abuse could have occurred and allowed for the proliferation of bacteria in food. The temperatures taken might not have been representative of the temperatures throughout different areas of the pans, because only a small number of temperatures were recorded relative to the number of pans of chicken served. It is likely that the temperatures varied within each pan of chicken and also between pans of chicken. During food preparation and serving, temperatures should be taken in different pans and in different locations within pans. It is especially important to take temperatures in the coldest areas (e.g. the pan farthest away from the heat source) to ensure that all of the food is consistently kept at a safe temperature. Given these factors, it is possible that portions of the food were held at temperatures that allowed for rapid bacterial growth.
It is also possible that the temperatures recorded did not capture the true variability of the temperature of the food over time. T he 10:15 am temperature recording would have been taken as the chicken was placed into the hot holding cabinet after cooking, and the noon temperature recording would have been taken about the time that the food was served. The temperature inside of the hot holding cabinet could have varied in an unpredictable way between these time points, given that transportation of the food required unplugging from an electrical source and the use of Sternos to maintain temperature. It is possible that the temperature varied between when it was plugged in and when Sternos were the only heat source and this variation was not captured when temperatures were taken at only two time points.
It is also possible that the temperature of the food decreased after the last recorded temperature, given that the buffet was open for another hour. This possibility is corroborated by several attendees who reported that the food, specifically the chicken, was cold. Our survey did not ask what time attendees ate or from which buffet line, so it is not possible to assess whether a particular line or time frame was associated with receiving cold chicken or with increased risk of illness. Finally, it is also possible that some attendees let food sit on their plates before eating it, and it could have cooled off that way.
The relatively short time that elapsed between cooking and consumption (3 hours from the time food was placed in the hot holding cabinet until the buffet closed) should have been adequate to prevent the significant growth of most, but not all, foodborne pathogens even if high enough temperatures were not maintained throughout all portions of the food for the entire time between cooking and service.
However, it is possible that one of the ingredients, including the pre‐cooked chicken, demi‐glace, mushrooms, or garlic was heavily contaminated before cooking, resulting in a high number of spores in the food with the potential for bacterial growth after cooking. We tested a few products which might not have been the actual ingredients that were used for food preparation; hence, the levels of C. perfringens and B. cereus in the actual ingredients used to prepare the chicken Marsala are not known.
Regardless of the level of bacterial contamination in the ingredients, rapid growth of C. perfringens after cooking could have been possible if high enough temperatures were not continuously maintained in all portions of the food. Because no food from the buffet lines was available for testing, the concentration of bacteria in the food consumed by the cases cannot be determined.
LIMITATIONS
Both the event organizer and Caterer A were made aware of multiple suspect foodborne illness complaints while the conference was ongoing; however, neither Caterer A (a required reporter per
COMAR 10.06.01.04), nor the event organizer reported the illnesses to the local health department.
Once reports were made to Baltimore City, BCHD, EIS did not immediately view the initial illness reports and did not associate them with Conference A right away, which further delayed reporting the cluster to BCHD, ACD. The intervening length of time from exposure to recognition of the outbreak decreased the likelihood of determining an etiologic agent, implicating a contaminated food, collecting clinical specimens from ill attendees and testing available conference food. To prevent these delays for future outbreaks and to ensure related complaints of illness are recognized, required reporters should be reminded of their obligation to report. Since the conclusion of this investigation, required reporters involved in this outbreak investigation, namely Caterer A, have been reminded of their obligation to report; it was suggested to the event organizer, while they are not a required reporter, that early reporting of suspect outbreaks are beneficial to investigations. A redundant reporting system has been created whereby 311 reports are sent to both ACD, EIS and others in BCHD.
The human specimens available for laboratory testing were collected one week or more after the onsets of illness, outside of the optimal time period for testing. Because the number of C. perfringens and B. cereus bacteria excreted in stool declines rapidly, and success for meaningful culture results outside of the optimal collection period is low, DHMH and the other state public health laboratories did not perform culture for these organisms. Some of the specimens were collected in Cary‐Blair media and the
PET‐RPLA assay kit specifies that testing should be done on stool not preserved in media. The effect of the media on testing for those few specimens is not known. Also, since some time had passed between the onsets of illness and collection of specimens, the detection C. perfringens in some specimens did not confirm that the bacteria were there at the time the cases became ill and negative results for other specimens did not establish that C. perfringens were not present previously. It is not possible to determine if CPE producing C. perfringens was present at the time of illness and at what level. And although C. perfringens bacteria were detected in some of the stools that were tested, C. perfringens can also be found in the stool of people who have not been sick.
Food samples collected and tested at DHMH may have been from the same manufacturer and possibly the same lots, however, were not samples from the ingredients in the food served at the conference.
These ingredients were not subject to the same conditions, such as cooking and hot holding, as the food consumed by the attendees. These limitations for laboratory testing may have prevented the early identification of an agent for the majority of cases.
Finally, the effect of the disclosure of the link to the online survey and password and the subsequent closing and reopening of the survey is unknown. However, we did not identify any obvious widespread issues with the information collected. Names and email addresses of survey respondents were checked against a list of registrants, and it appeared that there was no widespread discrepancy between the names on the 2 lists, meaning the survey respondents are presumed to likely be all registered conference attendees. Some survey respondents did not provide full names, or may have used nick names and different email addresses than were used for registration, so it is not possible to verify that all respondents actually attended. It is possible that fewer conference attendees responded to the survey as a result of closing and reopening the survey; however, having 604 total respondents with information used in the analysis should have provided enough study power to identify an implicated exposure.
While we did not conclusively identify exactly what happened during this outbreak, we did identify numerous opportunities for improvement to reduce the risk of illness for future events and to ensure a more timely response in the event future outbreaks occur.
Recommendations for event A and all event organizers and Caterer A and all foodservice facilities:1.) Ensure that internal food temperatures are measured at the conclusion of cooking and during the hot holding process.
a. Temperatures should be taken while the food remains inside the hot holding cabinets at one hour intervals and from multiple locations of the food trays on different shelves.
b. Food handlers should record the range of temperatures (versus a single temperature) as observed on log sheets. Food must maintain 135°F at all times after cooking and prior to service.
c. Obtain representative (multiple sites, mix of locations on tray, such as center, corners, edges) temperature measurements of all food trays before serving time.
2.) Report immediately to management or the person in charge when any food temperatures are below the required holding temperatures.
3.) Corrective action, as specified in the facility’s approved HACCP plan, must be taken when food measures less than the 135°F critical limit.
4.) Maintain detailed temperature logs.
a. Retain detailed internal temperature logs of any cold and hot held food every hour for all locations and all serving lines;
b. Log both internal and external temperature readings for all refrigeration units every 2‐4 hours to ensure that potentially hazardous foods do not exceed regulated time and temperature requirements.
No comments:
Post a Comment