Multistate Outbreak of Salmonella Infections Associated with Frozen Pot Pies --- United States, 2007
On June 6, 2007, a cluster of four human Salmonella serotype I 4,5,12:i:-* infections sharing a pulsed-field gel electrophoresis (PFGE) pattern was identified by the Pennsylvania Department of Health and reported to PulseNet. Initial investigations conducted during June--September 2007 by state and local health departments in collaboration with CDC did not identify a source of infection. This report summarizes the results of subsequent investigations of the outbreak, which determined that 401 cases of salmonellosis occurred in 41 states during 2007, with 32% of ill persons hospitalized. A multistate case-control study conducted during October 3--13 indicated that illness was associated with consumption of Banquet® brand frozen, not-ready-to-eat pot pies (odds ratio = 23.6; p<0.001). Further investigation determined that 77% of patients who ate these pies cooked them in microwave ovens and that consumer confusion regarding microwaving instructions might have resulted in a failure to cook the product properly. A voluntary recall was issued by the manufacturer (ConAgra Foods Inc., Omaha, Nebraska) on October 11, 2007, for all nine brands of pot pies produced at the implicated plant (plant A). The outbreak strain was isolated from 13 samples of unopened Banquet pot pies collected from the homes of patients. This outbreak highlights the need to cook not-ready-to-eat frozen foods thoroughly; these products should be clearly labeled as requiring complete cooking, and cooking instructions should be validated to account for variability in microwave wattage and common misconceptions among consumers regarding the nature of not-ready-to-eat foods.
An outbreak case was defined as infection with a Salmonella strain with the specific outbreak PFGE pattern and illness onset during January 1--December 31, 2007. During this period, a total of 401 outbreak cases from 41 states were identified (Figure 1). The median age of patients was 20 years (range: 1 month--97 years); 50% of the patients were female. Date of illness onset, known for 336 patients, ranged from February 20 to December 11, 2007, with a peak in September. Of patients with available information, 144 (50%) of 289 had bloody diarrhea, and 108 (32%) of 338 were hospitalized.
State and local health departments in collaboration with the CDC began the outbreak investigation on June 26, 2007. Interviews conducted over several months revealed frequent consumption of various chicken and egg food items. On October 3, 2007, a case-control study was initiated to ascertain the specific source of infections. Cases included Salmonella infections with the outbreak PFGE pattern in a resident of a participating state aged >2 years with no ill contacts and illness onset during August 1--October 3. Households with persons that could serve as age-group and neighborhood-matched controls were identified by an Internet-based reverse-address telephone directory. Investigators sequentially called telephone numbers until one to three persons with no diarrhea during the previous 2 weeks in the same age group as the case-patient were willing to serve as controls (or had caregivers willing to respond on their behalf). The exposure periods queried were 1 week before illness onset for case-patients and 1 week before interview for controls. As part of this study, epidemiologists from the Minnesota Department of Health (MDH) interviewed four case-patients who resided in Minnesota using the Minnesota standardized foodborne illness report form, a routine MDH practice. By October 4, the Minnesota epidemiologists reported that all four case-patients had consumed Banquet pot pies during the week before illness onset. After MDH notified CDC's OutbreakNet team of this finding, specific questions regarding pot pie consumption were included in the case-control study.
Of 48 eligible case-patients, 35 were enrolled. Six of these 35 were excluded because of exposure to other persons with diarrhea. Of the remaining 29, at least one matched control was enrolled for 17 case-patients. Data collected on the 17 matched sets were analyzed using exact conditional logistic regression (1). Measures of association between exposures and illness were calculated, using maximum likelihood estimates when available and median unbiased estimates when maximum likelihood estimates did not exist in the presence of complete data separation (2).
Case-patients were significantly more likely than controls to have eaten a Banquet pot pie (71% versus 0%, exact matched odds ratio [mOR] = 23.6 [median unbiased estimate], 95% confidence interval [CI] = 3.8--infinity). Banquet turkey pot pies were the only variety of Banquet pot pies associated with illness (29% exposure among case-patients versus 0% exposure among controls, mOR = 9.2 [median unbiased estimate], CI = 1.2--infinity). None of the remaining 67 exposures evaluated were associated with illness.
Pot Pie Consumption by Case-Patients
After the case-control study was completed, attempts were made to interview as many of the total 401 case-patients as possible using standardized forms focused on frozen food exposures. Of 236 case-patients for whom pot pie exposure information was collected, 174 (74%) reported consuming a frozen not-ready-to-eat pot pie during the week before illness onset. Most of these patients could name the brand or brands of pot pie consumed: 155 (92%) ate Banquet, three (2%) ate Banquet or another brand produced on the same production line at plant A, eight (5%) reported eating either Banquet or a non-recalled brand (i.e., a brand not recalled by the manufacturer), and three (2%) ate a non-recalled brand. A similar frequency of plant A pot pie consumption was observed among case-patients with illness onsets during April 28--June 21 as for those with onsets during August 10--December 31; exposure information was limited for other periods (Figure 2).
Banquet pot pie microwave instructions might have been confusing because different parts of the package recommended different preparation times. Furthermore, instructions for microwaving time varied by wattage. Of 133 patients interviewed, 102 (77%) cooked pot pies in a microwave. Of 78 patients who used a home microwave, only 23 (29%) reported knowing the wattage. Of eight patients who used a microwave outside the home, one (13%) knew the wattage. Forty-eight (68%) of 71 who responded did not let pies stand the full recommended time after microwaving, and 16 (19%) of 84 cooked more than one pie simultaneously, indicating that many patients did not follow microwaving instructions.
Voluntary Recalls and Environmental Investigations
On October 8, 2007, plant A suspended production of pot pies. On October 9, CDC and the U.S. Department of Agriculture's Food Safety and Inspection Service posted recommendations advising consumers not to eat pot pies with a production code ending in "P9," signifying a poultry-containing pie produced in plant A, the only plant that produces Banquet pot pies. On October 10, the advisories were expanded to include pies with production codes ending in "Est1059," signifying a beef-containing pot pie produced in plant A, because many patients could not report the exact meat in the pot pies consumed and some reported potentially having consumed Banquet beef pot pies. On October 11, ConAgra Foods issued a voluntary recall of all nine brands of pot pies produced in plant A. Before resuming production of these products, the manufacturer amended labeling and instructions.§
The outbreak strain was isolated from 13 unopened Banquet pot pies collected from the homes of patients. All pies that tested positive contained turkey, and all had production dates of July 13 or 31, 2007. Two of these pies underwent separate testing of the crust and filling. The filling tested positive and the crust tested negative for both pies.
Plant A produces approximately 1 million pot pies daily. Banquet pot pies are distributed nationally and account for approximately 75% of the national value-priced pot pie market.¶ Plant A also is the sole producer of identical pot pies packaged under various store brands.
Reported by: S Meyer, MPH, K Smith, DVM, Minnesota Dept of Health. I Azzam, MD, R Sowadsky, MSPH, Nevada State Dept of Health. IT Williams, PhD, O Henao, PhD, T Nguyen, MPH, J Austin, S Van Duyne, MS, Div of Foodborne, Bacterial, and Mycotic Diseases, National Center for Zoonotic, Vector-Borne, and Enteric Diseases; R Mody, MD, EIS Officer, CDC.
In this outbreak, 401 cases of Salmonella I 4,5,12:i:- infection were reported. However, for every reported case of Salmonella infection, an estimated 38 additional cases are not detected or reported (3). Therefore, many more persons might have been ill as a result of this outbreak. Consumption of Banquet pot pies was associated with illness, and testing of Banquet turkey pot pies collected from patients' homes yielded the outbreak strain. Mass food distribution can lead to widely distributed outbreaks, underscoring the importance of coupling laboratory-based surveillance of foodborne infections at the molecular subtype level with interviewing of patients to detect, solve, and truncate outbreaks.
The initial evidence that Banquet pot pies were the outbreak vehicle was acquired by MDH through the routine practice of combining data from PFGE subtyping of all Salmonella isolates and rapid interviewing of all patients. At MDH, these interviews used detailed food exposure questions to obtain open-ended histories, brand names, and purchase locations. Cross-referencing exposures identified in initial interviews and using an iterative approach to reinterview patients about suspicious exposures led to rapid identification of the possible outbreak vehicle.
Frozen, not-ready-to-eat microwavable meals have been reported previously as vehicles in salmonellosis outbreaks. Raw chicken nuggets and chicken strips were associated with Salmonella infections in a 1998 Australian outbreak and in Canadian studies of sporadic infections performed in 2003 (4--7). Stuffed chicken products were implicated in five outbreaks in Minnesota during 1998, 2005, 2006, and 2008 (MDH, unpublished data, 2008). Consumer confusion regarding the raw or cooked nature of these products was documented in these reports; products were not clearly labeled as containing raw poultry ingredients, and they were breaded and prebrowned, leading to the perception that they were precooked (6,7; MDH, unpublished data, 2008).
This outbreak differs from previously reported outbreaks with frozen, not-ready-to-eat food vehicles in that all meat ingredients were intended to be precooked before leaving plant A. However, the pot pies associated with this outbreak had a raw flour crust and were not-ready-to-eat, which allows consumers to prepare the food item to the level of doneness they prefer but also requires consumers to ensure that minimum cooking temperatures are reached to control microbiologic hazards. Furthermore, because raw frozen poultry pastes used to make the liquid portion of the chicken and turkey pie fillings enter plant A, pies might have contained undercooked poultry or been cross-contaminated from these raw poultry pastes, which often harbor Salmonella. Despite an intensive investigation of plant A and its ingredient suppliers, the source of contamination remains unknown.
This outbreak identified labeling concerns. Specifically, recommended microwave cooking times on the pot pie packaging were based on wattage categories, but most patients were unaware of their microwave wattage. Because of the small size of the case control study, the investigation could not determine whether microwaving pot pies rather than cooking them in a conventional oven was a risk factor for illness. Twenty-three percent of case-patients who ate a pot pie reported cooking the pies in conventional ovens, so improper microwave cooking could not account for the entire outbreak. However, given the observed limited knowledge about microwave wattage and the frequency of deviating from microwaving instructions, microwaving likely did lead to inadequate cooking.
Inadequate microwave cooking was thought to be partly responsible for two previous outbreaks of Salmonella infections (8,9). Industry and regulators should consider examining the manufacturing processes for frozen not-ready-to-eat foods to determine the extent to which microwave cooking is safe for these products. Labeling and cooking instructions on not-ready-to-eat frozen foods should be clear to ensure that consumers are aware of health risks and to facilitate compliance with validated cooking methods. Clear and prominent listing of output wattage on microwave appliances might improve consumer adherence to manufacturer's cooking instructions. Consumers should follow cooking instructions specific for an oven's wattage.
This report is based, in part, on contributions by state and local health department officials; P White, DVM, Food Safety and Inspection Svc, US Dept of Agriculture; and M Hoekstra, PhD, and A Sheth, MD, Div of Foodborne, Bacterial, and Mycotic Diseases, National Center for Zoonotic, Vector-Borne, and Enteric Diseases, CDC.
- Mehta CR, Patel NR. Exact logistic regression: theory and examples. Stat Med 1995;14:2143--60.
- Heinze G. A comparative investigation of methods for logistic regression with separated or nearly separated data. Stat Med 2006;25:4216--26.
- Voetsch AC, Van Gilder TJ, Angulo FJ, et al. FoodNet estimate of the burden of illness caused by nontyphoidal Salmonella infections in the United States. Clin Infect Dis 2004;38(Suppl 3):S127--34.
- Bucher O, Holley RA, Ahmed R, et al. Occurrence and characterization of Salmonella from chicken nuggets, strips, and pelleted broiler feed. J Food Prot 2007;70:2251--8.
- Currie A, MacDougall L, Aramini J, Gaulin C, Ahmed R, Isaacs S. Frozen chicken nuggets and strips and eggs are leading risk factors for Salmonella Heidelberg infections in Canada. Epidemiol Infect 2005;133:809--16.
- Kenny B, Hall R, Cameron S. Consumer attitudes and behaviours---key risk factors in an outbreak of Salmonella Typhimurium phage type 12 infection sourced to chicken nuggets. Aust N Z J Public Health 1999;23:164--7.
- MacDougall L, Fyfe M, McIntyre L, et al. Frozen chicken nuggets and strips---a newly identified risk factor for Salmonella Heidelberg infection in British Columbia, Canada. J Food Prot 2004;67:1111--5.
- Evans MR, Parry SM, Ribeiro CD. Salmonella outbreak from microwave cooked food. Epidemiol Infect 1995;115:227--30.
- Gessner BD, Beller M. Protective effect of conventional cooking versus use of microwave ovens in an outbreak of salmonellosis. Am J Epidemiol 1994;139:903--9.
* These four isolates were originally reported as serotype Typhimurium but later were determined to be serotype I 4,5,12:i:-.
PulseNet is the national molecular subtyping network for foodborne infections.
§ A label in large font on the front of the box that read "Ready in 4 Minutes" was changed to "Microwavable -- Must be cooked thoroughly. See back for directions." On the back of the box, improvements to the microwave cooking instructions included statements 1) advising the consumer to cook only one pot pie at a time, 2) advising the consumer not to use less than a 1,100-watt microwave to cook the pot pies, and 3) indicating that internal temperature of the pies needs to reach 165°F (74°C) as measured by a food thermometer in several spots.
¶ Value-priced frozen pot pies typically cost approximately $0.50, whereas premium frozen pot pies typically cost approximately $3.50.Figure 1
Return to top.
Return to top.
All MMWR HTML versions of articles are electronic conversions from typeset documents.
This conversion might result in character translation or format errors in the HTML version.
Users are referred to the electronic PDF version (http://www.cdc.gov/mmwr)
and/or the original MMWR paper copy for printable versions of official text, figures, and tables.
An original paper copy of this issue can be obtained from the Superintendent of Documents, U.S.
Government Printing Office (GPO), Washington, DC 20402-9371;
telephone: (202) 512-1800. Contact GPO for current prices.
**Questions or messages regarding errors in formatting should be addressed to email@example.com.