Investigating Food borne Outbreaks in the 21st Century

Investigating Food-borne st Century Outbreaks in the 21 Leah Lind, MPH Epidemiology Research Associate Pennsylvania Department of Health The New Yorker Bureau of Epidemiology Bureau of Epidemiology Communicable Disease Epidemiology Environmental Epidemiology Community Epidemiology Infectious Disease HIV/AIDS Asthma Cancer Food-borne Occupational illness STD Lead exposure Vaccine Preventable Pollutant exposure Bioterrorism Agents Epidemiologists ♦ Public Health Physicians ♦ Toxicologists ♦ Support Staff What does an epidemiologist do exactly?   Epidemiology – the study of patterns, causes, and control of disease in groups of people Patterns - data and statistics   How many people in PA get Salmonella every year? Are giardiasis rates higher in one neighborhood? Smoking causes lung cancer Thalidomide during pregnancy causes severe birth defects Vaccinations Behavior modification  Causes - risk factors for diseases and conditions    Control – public health   Are giardiasis rates higher in one neighborhood?      Giardiasis – GI illness caused by Giardia lamblia, usually waterborne In PA about 8/100,000 people will get giardiasis each year (normal pattern) In July in one neighborhood of 200 people, 6 people have been diagnosed with giardiasis (abnormal pattern) Common water supply? Is water supply infected? (cause) Discontinue use of tap water until problem is resolved (control) Where does the DOH get their data?  Lab tests   Positive lab results for about 80 conditions are automatically reported to us through a computerized system Constitutes the bulk of disease reports to the DOH    Doctors’ offices/Hospitals/Clinics School nurses Self-report - rare What happens to the data?  Public health nurses   View new disease reports every day Perform investigation of disease and add more data   Data are analyzed to determine how many of each disease is reported each week, peaks and dips in numbers are investigated Data is reported to CDC each week, CDC then tallies national numbers http://www.cdc.gov/mmwr/ What is a disease investigation?     Public health nurse receives disease report May need to call doctor for more information on patient’s condition For all food-borne disease investigations the nurse calls the patient directly Will ask patient a series of questions specific to the disease Food-borne Disease Investigation        Symptoms and onset (diarrhea, nausea, vomiting, fever, abdominal pain, achiness) Travel Water source Exposure to animals Swimming Restaurants visited All foods eaten in days prior to illness Why are the numbers high?  Seasonal fluctuation?    Summer is peak time for food-borne illness Temperature violations common Frequent travel   Coincidence? Statistical interpretation?   Salmonellosis cases doubled this week!!! Actuality – salmonellosis cases went from 2 to 4 this week  Outbreak? What is a food-borne outbreak?    CDC definition: An outbreak of food-borne illness is a cluster of 2 or more infections caused by the same agent (pathogen or toxin) which upon investigation are linked to the same food. 2 cases of Salmonella in the same household who both ate chicken for dinner 2 nights ago? Outbreak 600 cases of E. coli in persons who visited same hamburger restaurant? Outbreak Causes of food-borne illness  Bacterial       Botulism Campylobacterosis E. coli Listeriosis Salmonellosis Shigellosis Causes of food-borne illness  Parasitic    Cryptosporidiosis Cyclosporiasis Giardiasis  Viral   Hepatitis A Norovirus Botulism        Caused by Clostridium botulinum Very rare, about 110 cases in US each year One of top 5 bioterrorism agents Potentially fatal Can cause total paralysis of nervous system leading to stoppage of heart Often due to home canned foods, infant botulism often due to honey Avoid home canned foods, children under 12 months should not be fed honey Campylobacteriosis      Caused by Campylobacter Diarrhea, abdominal pain, cramping about 2-5 days after eating infected food One of the most common foodborne illnesses, about 1500 cases in PA each year Usually do to raw/undercooked poultry, can also be due to water infected from bird droppings Standard food-borne illness prevention techniques E. coli    E. coli O157:H7 is most common cause of E. coli illness Bloody diarrhea and abdominal cramps beginning 2-8 days after eating contaminated food Many sources identified    Most common – undercooked beef Spinach, sprouts, lettuce have been implicated in recent outbreaks Swimming in/drinking sewage contaminated water  Standard food-borne illness prevention techniques Listeriosis    Caused by Listeria, typically Listeria monocytogenes Fever, muscle aches, sometimes nausea and vomiting Most healthy persons exhibit no symptoms of listeriosis; pregnant women and immunocompromised are most at risk   May cause miscarriage in pregnant women Can cause serious illness and possible death in elderly, cancer, diabetes and AIDS patients    About 50 cases/year in PA Often due to soft cheeses, hot dogs and deli meats At risk persons should avoid soft cheeses and unpasteurized milk, heat hot dogs and deli meats thoroughly before eating Salmonellosis       Caused by one of many varieties of Salmonella Diarrhea, abdominal pain, fever about 13 days after eating infected food Common cause of food-borne illness, about 1500-2000 cases/year in PA Most commonly in undercooked eggs, poultry and meat but can be found in almost any food including many fresh produce, has even been found in cereal Recent outbreaks have been due to tomatoes Standard food-borne illness prevention techniques, use pasteurized eggs Shigellosis       Caused by Shigella, typically Shigella sonnei Diarrhea, fever, stomach cramps 1-2 days after eating contaminated food About 100-300 cases/year in PA Most common mode of transmission is poor hand washing hygiene after using bathroom and prior to preparing food, however, eating vegetables grown in fields with contaminated water and drinking contaminated water can cause infection Most outbreaks occur in daycare centers Practice proper hand hygiene, wash fruits and vegetables and drink only treated water Cryptosporiasis      Caused by Cryptosporidium parvum and hominis Watery diarrhea, stomach cramps, nausea, vomiting 210 days after eating contaminated food About 150-200 cases/year in PA Usual source is contaminated water; raw fruits and vegetables can also be contaminated Wash and peel all vegetables to be eaten raw, drink only treated water Cyclosporiasis      Caused by Cyclospora cayetanensis Watery diarrhea, stomach cramps, nausea, vomiting about 7 days after eating contaminated food Rare in the US, about 5 cases/year in PA Usually found in fresh produce imported from developing countries Large outbreak traced to Guatemalan raspberries in 1990s Giardiasis       Caused by Giardia lamblia Diarrhea, stomach cramps, nausea 7-14 days after exposure Over 1000 cases in PA/year Usually found in contaminated water; can also be found in raw produce Most outbreaks are waterborne Wash and peel all vegetables to be eaten raw, drink only treated water Hepatitis A      Jaundice, abdominal pain, nausea, diarrhea and fever 15-50 days after exposure About 200 cases/year in PA Usually due to produce, typically imported from endemic countries, and infected food workers Recent outbreaks attributed to strawberries and green onion Vaccine is available for persons known to be exposed Norovirus     Nausea, vomiting, diarrhea, abdominal cramps about 12 days after exposure Most common GI illness Can be transmitted through contaminated food and person-to-person Practice proper hand hygiene, wash fruits and vegetables and drink only treated water Investigating an Outbreak the Old Fashioned Way     Once multiple cases of illness have been established, we look for a common meal During the standard food-borne illness investigation by the public health nurse, she will ask if the person attended a large gathering If yes, the person will be asked if other people from that gathering are ill If yes, we identify that meal as the source of the outbreak Investigating an Outbreak the Old Fashioned Way   We ask the patient for a contact person for the meal From the contact person we request    The full menu from the meal Names and contact numbers for all attendees Contact information for the caterer, restaurant or person(s) who prepared the meal Investigating an Outbreak the Old Fashioned Way  Menu  We develop a questionnaire specific to the outbreak with questions on symptoms, demographics and each food item served at the event  Names and contact numbers for all attendees   Using the questionnaire we contact each person in a small outbreak or a subset in a large outbreak and administer the questionnaire – we contact both ill and well attendees Encourage ill persons to submit stool samples to identify source of outbreak Investigating an Outbreak the Old Fashioned Way  Contacting the food preparers  We will contact the Department of Agriculture or the local health officer to perform an inspection of the caterer or restaurant (in PA, the DOH does not perform inspections, only Ag or if the local health officer has jurisdiction to perform the inspections)  The Ag inspector will perform an inspection and ask a series of questions about the menu       How long was the meat cooked? What was the internal temperature? Was it left unrefrigerated? Was the fresh produce washed? Where were foods purchased? What were the ingredients in the stuffing?  They may also take food and environmental samples for testing Investigating an Outbreak the Old Fashioned Way   Once data from questionnaires is collected, it is entered into a database and analyzed Results are presented in statistical terms  p-value = on a scale of 0 to 1, the likelihood that the findings are due to chance with 1 being the most likely and 0 being the least likely; the least likely food item is generally identified as the culprit  Odds Ratio (OR) = the odds of becoming ill if you ate a food item     OR 1 means a 1:1 ratio, so you’re no more likely OR >1 means >1:1 ratio, so you’re more likely to become ill from having eaten this food item OR<1 means <1:1 ratio, so you’re more likely to have remained well if you ate this food item Therefore, the food item with the highest OR is generally identified as the culprit Investigating an Outbreak the Old Fashioned Way  Example of statistical results Menu Item p-value Hot dogs 0.56 OR 1.04 Potato salad Beans 0.34 0.04 Lowest p-value 0.85 3.11 Highest odds ratio Investigating an Outbreak the Old Fashioned Way   Does it make sense? What do we know about the beans? Further investigation reveals:     Beans cooked with pork Temperature of cooked beans not checked Beans left on counter after cooking, not refrigerated or kept hot Beans transported one hour in hot car  Conclusions – beans are most likely source of food poisoning Real Life Outbreak #1 - Cylclospora      Parasite, first reported in 1979 Reporting became more frequent in 1980’s, most cases reported in 1990’s About 150 infections annually, primarily in travelers Usually found in fresh produce – raspberries, fresh basil Largest U.S. outbreak: contaminated raspberries from Guatemala in 1996 (978 cases, no deaths) Real Life Outbreak #1 - Cylclospora    Setting  June 2004  Quiet day at the office on a Friday afternoon  State Health Center nurse calls to inform us that local hospital just had their 3rd Cyclospora case that week Is this unusual?  In PA about 5 cases/year of Cyclospora  3 cases in one week at one local hospital – very unusual Initial information  According to nurse, these 3 persons had all attended a luncheon at a local nursing home  Nursing home was contacted, menu and list of attendees provided to DOH  Nursing home also informs DOH that a number of residents have become ill with diarrheal illness and stools are being sent out Real Life Outbreak #1 - Cylclospora  Nursing home was able to identify 5 meals in which participants became ill in the days following      Meal #1 - late May Meal #2 – early June Meal #3 – late June Meal #4 – late June Meal #5 – late June Real Life Outbreak #1 - Cylclospora      Department of Agriculture was called to investigate facility Menus from each of these events were provided to DOH Attendee lists for all events were provided to the DOH Questionnaires were developed for both ill and well attendees on symptoms, events attended, food items eaten During the course of interviewing, more persons were becoming ill and more stool sample results were coming in Real Life Outbreak #1 - Cylclospora  Results    96 were identified as ill out of 215 persons who attended the various events (45%) 39 were confirmed by lab tests showing cyclospora in the stool DOA found no significant deficiencies from inspection  Food results – food item at each event with lowest p-value and highest OR      Meal #1 – pasta salad (p=0.0005,OR=11) Meal #2 – pasta salad (p=0.04,OR=1.3) Meal #3 – pasta salad (p=0.0003,OR=23) Meal #4 – pasta salad (p=0.0001,OR=4) Meal #5 – pasta salad (p=0.0001,OR=11) Real Life Outbreak #1 - Cylclospora    What in pasta salad? Prepared with pasta, salad dressing, dried herbs and fresh vegetables including tomatoes, carrots, snow peas, cabbage Most likely one of the fresh vegetables; cook informed us he had used different bags or shipments of vegetables for each vegetable except snow peas All vegetables had been thoroughly washed, however, Cyclospora is particularly difficult to remove from produce as it is encapsulated in a spore which is better able to cling to its host and makes it more resistant to killing Real Life Outbreak #1 - Cylclospora      Same bag of snow peas used for each batch of pasta salad Snow peas were only used in pasta salad, were not served at any other meals, but the other vegetables were served at meals where no one became ill Snow peas were traced to Guatemala Guatemalan raspberries were implicated in a large outbreak in the 90’s after which US import of Guatemalan raspberries was temporarily suspended Published in MMWR http://www.cdc.gov/mmwr/preview/mmwrhtml/mm53 e917a1.htm Real Life Outbreak #1 - Cylclospora Real Life Outbreak #2 – Hepatitis A      Viral illness, unlike Hepatitis B and C, can be transmitted via food About 5,000 cases reported a year, previously much higher rates but increases in vaccinations have lowered rates Can be passed person-to-person and fecal contamination of foods can occur in endemic countries due to poor sanitation in fields Endemic areas include Mexico, African countries, Southern Asian countries Areas with low rates of disease include US, Canada, Western Europe, Australia Real Life Outbreak #2 – Hepatitis A National Hepatitis A Counts 35000 30000 25000 20000 15000 10000 5000 0 19 93 19 94 19 95 19 96 19 97 19 98 19 99 20 00 20 01 20 02 20 03 20 04 Year Real Life Outbreak #2 – Hepatitis A  Setting    November 2003 Saturday ED physician in Beaver county noted 6+ cases of Hepatitis A in past week, notified on call public health physician  Unusual?   About 200 cases/year in PA, so not shocking 6 in one week at one hospital is significant for this area, however Interviews all cited Restaurant A in Southwestern PA  Initial Information  Real Life Outbreak #2 – Hepatitis A      First – a note about confidentiality The PA DOH takes confidentiality very seriously We won’t release names of patients even as part of contact investigations We won’t release names or facility names to press or public unless we feel the facility poses an ongoing threat That said, if the facility name is released by the press through their investigation, we are at liberty to mention the facility name Real Life Outbreak #2 – Hepatitis A     Restaurant A was identified as Chi Chi’s in Beaver county Chi Chi’s had declared bankruptcy in October 2003, the month preceding the outbreak Chi Chi’s had 65 locations in the NE/midwest, only this location associated with illness Department of Agriculture called to perform inspection Real Life Outbreak #2 – Hepatitis A     Hepatitis A is most frequently spread via a fecal-oral route May be spread up to 2 weeks before symptoms begin to appear Food workers are often source of Hep A transmission due to poor hand washing/gloving techniques Vaccine is available, but not federally required for food workers (some local governments do require vaccination for food workers) Real Life Outbreak #2 – Hepatitis A  DOH Response     Food workers at Chi Chi’s were questioned and possible Hep A was identified in several who were subsequently tested for Hep A 200-600 persons served a day at this location Mass immunization clinics were held for persons who had eaten at Chi Chi’s in the previous 2 weeks Customers were questioned regarding the food items consumed at Chi Chi’s Real Life Outbreak #2 – Hepatitis A  Results      527 cases in PA >440 cases in Beaver County 74 cases from other states >9400 immunizations given 3 deaths, 1 survivor had liver transplant Initial food questionnaires suggested mild salsa Analysis of ill customers vs well customers showed mild salsa was most likely culprit (OR=20) Specifically, green onions were shown to be the most likely culprit (OR=33); green onions were in the salsa and also garnish for some other dishes  Food results    Real Life Outbreak #2 – Hepatitis A      Initially thought to be caused by food workers at Chi Chi’s, however, those who tested positive became ill after eating the green onions Green onions were traced to a farm in Mexico where Hepatitis A is common Contamination at farm was most likely source of Hepatitis A Contaminated green onions only made it into the bag sent to Beaver County Chi Chi’s The largest “single source” outbreak in the US Real Life Outbreak #2 – Hepatitis A PA DOH Laboratory     Laboratory science has come a long way Once a pathogen is identified at a local lab (eg. Salmonella, E. coli, etc.) it is then sent to the PA DOH lab for further testing Pennsylvania Bureau of Laboratories (BOL) is located in Southeast PA BOL has 2 mobile labs which can be deployed for outbreaks, one located in eastern PA, one in western PA PA Mobile Labs Laboratory Testing    New laboratory techniques allow us to identify pathogens more specifically Serotyping can be done  Bacteria often have large groups with several different types  BOL can test to determine what the type is  Eg. Salmonella has 2500 different types such as Salmonella enteritidis, Salmonella typhimurium, Salmonella javiana, etc.  E. coli is actually a specific type of Escherichia bacteria called Escherichia coli  A common type of Campylobacter is Campylobacter jejuni Serotyping can help us determine if different cases of a bacteria are related Serotyping    Recent peanut butter salmonella outbreak was found to be Salmonella tennessee Since peanut butter is such a common food item, it was very helpful to eliminate anyone from the outbreak who had had peanut butter, but had a different type of Salmonella such as S. enteritidis or S. javiana Had we not been able to narrow down the salmonella cases who consumed peanut butter to only those with S. tennessee, our outbreak numbers would have been skewed very high PFGE     Once serotyping has been employed, the lab will use a technique called Pulse Field Gel Electrophoresis (PFGE) Provides a genetic fingerprint of the bacteria so we can further narrow down persons associated with an outbreak S. typhimurium is a very common type of salmonella S. typhimurium JPXX01.022 is quite rare PFGE   Can be applied to any bacteria CDC maintains a database of all PFGE patterns and names new ones PulseNet   Letters indicate type of bacteria, numbers are assigned in the order in which they are identified So JPXX01.0022 and JPXX01.0023 may not look anything alike, but they happened to be identified sequentially  We can compare our PFGE patterns to patterns seen throughout the US to find that outbreaks may be occurring nationwide, not just in PA PFGE PFGE-XbaI PFGE-BlnI 06E02106 06E02122 06E02123 32 58 35 Do all of these PFGE patterns match? 06E02124 06E02142 06E02149 06E02188 07E00026 07E00033 07E00035 07E00151 43 64 11 50 41 22 52 22 PFGE PFGE-XbaI PFGE-BlnI 06E02106 06E02122 06E02123 32 58 35 These 3 do not 06E02124 patch the other 06E02142 patterns, but they 06E02149 are very close 06E02188 07E00026 07E00033 07E00035 07E00151 43 64 11 50 41 22 52 22 PFGE Use in Outbreaks    PFGE allows us to find seemingly unrelated cases to identify common exposures Allows us to include or exclude persons in an outbreak, especially with a common type If a food item is available for testing, we can match the PFGE pattern of the bacteria in the food to the PFGE pattern in the stool of the ill persons Real Life Outbreak #3 - Salmonella      Bacterial infection About 40,000 cases reported a year in US Mild cases generally not reported, actual rate of Salmonella infection estimated to be 30x reported rate (~1,200,000) Salmonella enteritidis primarily found in eggs; also found more often in raw meat, raw milk; can be found in almost any food product Can also be transmitted from fecal matter of pet birds and reptiles Real Life Outbreak #3 - Salmonella  Setting    July 2004 PA BOL begins to identify Salmonella javiana, a previously uncommon Salmonella type and alerts Epi S. javiana counts identified at PA BOL in 2004 prior to July Month Number January February 1 1 March April 1 0 May June 1 0 Real Life Outbreak #3 - Salmonella  S. javiana seen in the first 2 weeks of July 2004 at PA BOL Week Number July 1-7 July 8-14 2 41 Real Life Outbreak #3 - Salmonella First PFGE testing done on S. javianas, 9 are matches Real Life Outbreak #3 - Salmonella     A previously uncommon type of Salmonella and PFGE pattern are now showing a large surge in July 2004, almost certainly an outbreak Most cases are in Western PA and cases did not attend a common event Initial interviews show a common exposure among these otherwise unrelated persons – Sheetz deli Initial study indicates tomatoes or lettuce may be the culprit Real Life Outbreak #3 - Salmonella      The outbreak expanded to include central PA and states in surrounding areas Other types and PFGE patterns of salmonella were found to be involved As the outbreak progressed it was found that other types and PFGE patterns of Salmonella were involved in the outbreak Roma tomatoes were identified as the most likely cause of the outbreak Roma tomatoes tested at the lab revealed one of the less common serotypes/PFGE patterns involved in the outbreak Real Life Outbreak #3 - Salmonella Real Life Outbreak #3 - Salmonella     Eventually 429 cases from 9 states were linked to the outbreak on the basis of serotyping/PFGE patterns and having eaten at a Sheetz deli prior to becoming ill Before these lab advances this outbreak may not have even been picked up – increases in Salmonella in the summer is very common Due to lab advances we were able to link persons in multiple states to one outbreak – prior to this time, it was difficult to determine if cases were linked to an outbreak within a state let alone multiple states A trace back of the tomatoes led to tomatoes grown in Florida Real Life Outbreak #4 – E. coli O157:H7     Shigatoxin producing E. coli 73,000 infections annually (1999 CDC estimate)  at least 61 deaths Was primarily seen in uncooked beef, now becoming more common in produce Largest U.S. outbreak: contaminated well water at NY county fair, 1999 (781 cases, 2 deaths) Real Life Outbreak #4 – E. coli O157:H7  Setting    September 2006 OR, NM, WI investigating increases in E. coli O157:H7 Oregon suspects fresh spinach   PFGE patterns involved in this outbreak are posted on PulseNet It is quickly noticed that this PFGE pattern is being seen in multiple states in the US, not just the 3 investigating outbreaks – 8 states are believed to be involved Real Life Outbreak #4 – E. coli O157:H7   PFGE pattern identified in outbreak E. coli EXHX01.0124 PFGE-XbaI PFGE-BlnI 06E01397 06E00885 06E01455 06E01456 06E01457 06E01458 06E01464 06E01465 06E01510 20 3 24 19 12 64 25 2 58 F F F F M F F F F Real Life Outbreak #4 – E. coli O157:H7 Annual Counts of E. coli EXHX01.0124 Nationwide 250 200 150 100 50 37 233 0 1998 1999 2000 2001 2002 2003 2004 2005 2006 Real Life Outbreak #4 – E. coli O157:H7     Eventually, 205 cases were linked to the outbreak nationwide due to matching PFGE patterns, including 3 deaths PA had 10 cases linked to the outbreak with a matching PFGE pattern, no deaths Spinach was identified early on by Oregon as the culprit, as the outbreak grew nationally, questionnaire data specifically identified bagged spinach sold in grocery stores as the culprit FDA press releases advised all to avoid fresh bagged spinach Real Life Outbreak #4 – E. coli O157:H7 Ontario 50 11 10 26 18 26 US states and 1 Canadian province were found to be involved in outbreak Real Life Outbreak #4 – E. coli O157:H7     13 bags of spinach from ill consumers were tested All 13 were found to have E. coli with a PFGE pattern matching the outbreak pattern FDA trace back revealed spinach was grown on farms in California Natural Selection Foods voluntarily recalled bagged spinach Real Life Outbreak #4 – E. coli O157:H7 E. Coli O157:H7        Late 2006 another outbreak of E. coli O157:H7 Involved northeastern US states – PA, NY, NJ, DE New Jersey first linked cases to Taco Bell Initial reports by Taco Bell indicated green onions were the culprit due to lab testing However, results could not be duplicated Investigation pointed to lettuce as the most likely cause of illness Lettuce traced back to farms in California E. Coli O157:H7       Also in late 2006 another outbreak of E. coli O157:H7 Involved midwestern US states – MN, IA Linked cases to Taco John’s restaurant Unrelated to Taco Bell outbreak in the NE Investigation pointed to lettuce as the most likely cause of illness Lettuce traced back to farms in California Norovirus      “Noro” is a very common cause of diarrheal and vomiting illness Most common outbreak investigated by PA DOH Person-to-person spread very common, also often spread through food Noro outbreaks occur in long term care facilities, usually these are person-to-person However, these can be spread by ill food workers, so ill food workers should not be handling food while ill Moral of the Stories  Case questionnaires in food-borne illness are of vital importance   These are usually done before we receive PFGE results, so if multiple cases are citing the same restaurant there may be something there After lab results indicate an outbreak may be occurring we can review questionnaires to get an idea of the source   The advent of serotyping and PFGE has improved not just our ability to “catch” outbreaks, but to better define an outbreak and who’s involved Epidemiology is the link between questionnaire and lab data Don’t Panic      US food supply is one of the safest in the world Many laws to oversee food production and distribution Food-borne illness surveillance and investigation occurs in all states All states are required to report cases of food-borne illness to CDC Many state and federal agencies involved in effort to maintain high level of food safety   FDA – Center for Food Safety and Applied Nutrition (CFSAN) Roles of all agencies involved in food safety: http://www.cfsan.fda.gov/~lrd/foodteam.html Resources  CDC - www.cdc.gov   MMWR - http://www.cdc.gov/mmwr/ Food-related Diseases http://www.cdc.gov/ncidod/diseases/food/index.ht m    PA Department of Health – www.health.pa.us FDA – www.fda.gov USDA – www.usda.gov http://www.fightbac.org/