«Office of Compliance Center for Food Safety and Applied Nutrition July 21, 2016 CONTENTS EXECUTIVE SUMMARY BACKGROUND OBJECTIVES SAMPLE COLLECTION ...»
FY 2014 – 2016 Microbiological Sampling Assignment
Summary Report: Raw Milk Cheese Aged 60 Days
Office of Compliance
Center for Food Safety and Applied Nutrition
July 21, 2016
Pathogen Findings: Salmonella
Pathogen Findings: Listeria monocytogenes
Pathogen Findings: E. coli O157:H7
Pathogen Findings: Shiga toxin-producing E. coli
Pathogen Findings: By Country of Origin
Pathogen Findings: By Repeat Firm with Adulterated Samples
OTHER FINDINGS: GENERIC E. COLI
PUBLIC HEALTH IMPACT
Voluntary Industry Actions and Regulatory Activities for Domestic Sample Pathogen Findings
Voluntary Industry Actions and Regulatory Activities for Imported Sample Pathogen Findings
Voluntary Industry Actions and Regulatory Activities for Generic E. coli Findings
CONCLUSION AND NEXT STEPS
APPENDIX A: TEST METHODS
APPENDIX B: POSTIVE FINDINGS BY BACTERIAL TYPE
APPENDIX C: GENERIC E. COLI TESTING SCHEME AND REGULATORY STRATEGY
APPENDIX D: GENETIC EVALUATION
The new approach, detailed in the Background section of this report (page 4), centers on the testing of a statistically determined number of samples of targeted foods over a relatively short period of time, 12 to18 months, to ensure a statistically valid amount of data is available for decision making. This approach helps the agency determine if there are common factors – such as origin, variety or manufacturing practice – associated with any pathogen findings.
The FDA issued the raw milk cheese assignment in January 2014 along with two others (for sprouts and avocados) as the initial commodities under its new sampling model. As planned, the FDA collected 1,606 raw milk cheese samples (exceeding its target by 6 samples). The FDA designed its sampling plan such that if contamination of one percent or greater was present in the commodity, the agency would detect it. The agency closely monitored the assignment to gather lessons learned and make changes to the sampling if needed to address trends or food safety issues.
Of the 1,606 raw milk cheese samples collected and tested, 473 samples (29 percent) were domestic samples, and 1,133 samples (71 percent) were of international origin. The FDA sought to design its sampling plan to approximate the ratio of domestically made versus imported product on the U.S. market but was unable to do so in this case because the federal government does not track production volume of raw milk cheese. 1 Details on the assignment design are provided in the Sample Collection section of this report (page 6).
The FDA tested samples for the presence of the pathogens Salmonella, Listeria monocytogenes, E. coli O157:H7 and Shiga toxin-producing E. coli, as well as for generic E. coli. The overall contamination rate for each of the pathogens was less than one percent, and the overall contamination rate for generic E. coli was 5.4 percent. While the prevalence for generic E. coli was comparatively high, it bears mention that it rarely causes illness even as it may signal insanitary processing conditions.
Because the contamination frequencies among the pathogens were below one percent, the FDA was limited in its ability to detect differences in contamination rates based on the type of cheese or its origins (i.e., domestic vs. import), even with the large number of samples.
In addressing the violative samples of domestic raw milk cheese, the FDA worked with the responsible firms to carry out recalls as appropriate and followed up with facility inspections. In addressing the violative samples of imported raw milk cheese, the FDA refused entries of raw 1 The USDA Economic Research Service tracks the supply and commercial use of cheese in the United States but has no figures specific to raw milk cheese, which differs mainly with respect to the lack of pasteurization even as it comes in many varieties.
Listeria monocytogenes in cheeses, particularly semi-soft varieties, remains a concern, as demonstrated by the nearly one percent contamination rate in semi-soft cheeses (see Appendix B: Positive Findings by Bacterial Type). The FDA believes this contamination rate may be related to product handling practices or procedures. Given the serious public health implications of Listeria monocytogenes contamination associated with ready-to-eat foods, the FDA plans to continue to work with the cheese industry to identify and correct practices that lead to Listeria monocytogenes contamination in cheese.
The FDA Food Safety Modernization Act is an amendment to the Federal Food, Drug, and Cosmetic Act (FD&C Act) that is intended to provide the FDA with additional authority to better prevent problems before they occur. However, in order to develop prevention-based systems, the FDA needs data and other information to help identify hazards that must be addressed and minimized. That is why sampling is an important part of the agency’s preventive approach to food safety and why the FDA has developed a new microbiological sampling model designed to identify patterns that may help predict and prevent contamination by disease-causing bacteria.
The new model complements the FDA’s longstanding approach to sampling, which has employed for-cause and targeted strategies to monitor known hazards. The FDA will continue its longstanding approach to sampling, but the prevention mandate outlined in FSMA gives cause for larger, in-depth surveys of products and commodities to better evaluate risks. These large sample collections enable the FDA to determine the prevalence of contamination in instances where it does not otherwise have enough data to do so. Such studies also may shed light on areas of needed focus or issues of food safety that must be addressed. However, if a contamination rate is substantially lower than one percent, then even extensive sampling may not provide actionable insight into the source of the contamination.
As a starting point for the new model – and because it is not feasible to sample every product and/or commodity extensively – an FDA work group developed a system to rank commodities based on microbial risk. The work group reviewed sampling data collected over a five-year period, systematically considering criteria such as association with foodborne illness, consumption of product without a mitigating “kill” step, and prevalence data (i.e., the number of samples that tested positive for a pathogen in proportion to the total number of samples tested for the given commodity). Foods that ranked comparatively high were evaluated by subject matter experts to determine their feasibility as candidates for a large-scale survey and the remaining data needs for the commodity. After the work group review, the FDA chose to sample raw milk cheese (aged 60 days), sprouts (seeds, spent irrigation water, and finished product), and avocados (whole pit fruit) in FY2014-2016, as the first commodities under the new model. This report details the rationale and findings for the sampling and testing of raw milk cheese.
4Why Raw Milk Cheese?
Evidence indicates that aging raw milk cheese for 60 days may not eliminate or adequately reduce E. coli O157:H7 and Salmonella in raw milk cheese, thus posing a potential hazard to consumers. 2, 3 Additionally, Listeria monocytogenes may grow in certain types of cheese during aging. 4 The Centers for Disease Control and Prevention reported in a 2012 study that cheese made from raw milk was involved in 27 outbreaks of foodborne illness from 1993 to 2006. 5
Raw Milk Cheese Production
The USDA Economic Research Service (ERS) tracks the supply and commercial use of cheese in the United States but has no figures specific to cheese made from unpasteurized milk. In 2014 the United States produced more than 11 billion pounds of cheese and imported 363 million pounds of the product. 6,7 The same year, the most recent for which statistics are available, USDA ERS also reported that Americans ate an average of 33.9 pounds of cheese per person. That figure has more than doubled since 1975 when U.S. cheese consumption was just 14.2 pounds per person. 8 Based on anecdotal information, the FDA believes that the volume of unpasteurized cheese on the U.S. market may also have increased over this time period, in part because of the apparent growth in specialty cheese manufacturing. Despite the lack of quantitative data on raw milk cheese, the increase in cheese consumption in the United States suggests that consumption of raw milk cheese may, too, be on the rise.
Making raw milk cheese entails the same basic steps as are required for any cheese, whether made by an artisan or a comparatively large manufacturer. Typically, the cheesemaker receives raw milk for processing, heats it and adds starter culture, followed by rennet, then allows it to set (or coagulate). Once set, the curd is cut to release the whey. The mixture is cooked and stirred until the desired temperature and firmness are reached, and the whey is drained. The 2 Reitsma, C.J., Henning, D.R. (1996). Survival of Enterohemorrhagic Escherichia coli O157:H7 during the Manufacture and Curing of Cheddar Cheese. Journal of Food Protection, Vol. 59, No. 5, Pages 460-464.
3 Schlesser, J.E., Gerdes, R., Ravishankar, S., Madsen, K., Mowbray, J. & TEO, A.Y. (2006). Survival of a Five-Strain Cocktail of Escherichia coli O157:H7 During the 60-Day Aging Period of Cheddar Cheese Made from Unpasteurized Milk. Journal of Food Protection. Vol. 69, No. 5, Pages 990-998.
4 Joint FDA / Health Canada Quantitative Assessment of the Risk of Listeriosis from Soft-Ripened Cheese Consumption in the United States and Canada.
(http://www.fda.gov/Food/FoodScienceResearch/RiskSafetyAssessment/ucm429410.htm) 5 Langer, A.J., Ayers, T., Grass, J., Lynch, M., Angulo, F., & Mahon, B.E. (2012). Non-pasteurized Dairy Products, Disease Outbreaks, and State Laws—United States, 1993–2006. Emerging Infectious Diseases. Vol. 18, No. 3, Pages 385-391.
6 USDA Dairy Products 2015 Summary. (http://usda.mannlib.cornell.edu/usda/current/DairProdSu/DairProdSu-04pdf) See page No. 30, “Total Cheese (Excluding Cottage Cheese) Production by Month – States, United States, and Regions: 2015 and Total 2014-2015.” 7 USDA Data Set: Value of U.S. Dairy Products.
(http://www.ers.usda.gov/datafiles/US_Food_Imports/__Value_of_US_food_imports_detailed_tables_by_food_gr oup/dairy2_1_.xlsx) 8 Dairy products: Per capita consumption, United States, 1975-2014.
Worldwide there are more than 2,000 cheese varieties. Virtually all of them fall within the basic categories of “soft (fresh),” “soft-ripened,” “semi-soft,” “hard,” or “hard grating.” But, cheese is a broad category of food that features many variables. For example, it may be made from the milk of cows, goats, or sheep, or a combination thereof, or even other animals. It may be moldripened, have a washed rind, or a natural rind. It may have a closed texture, like Cheddar, or have an open texture, like Swiss. And, it may be made from pasteurized, thermized or raw milk.
The objectives of the FY2014-2016 raw milk cheese sampling assignment were:
• To determine the prevalence of Salmonella, Listeria monocytogenes and Shiga toxinproducing E. coli (primarily E. coli O157:H7) in raw milk cheese aged 60 days.
• To determine if there are common factors associated with positive findings (such as origin, variety or manufacturing practice).
• To take appropriate regulatory action when positive findings are observed.
• To explore new processes and parameters to strengthen the FDA’s current approach to sample collection and analysis.
SAMPLE COLLECTIONThe FDA collected 1,606 samples of raw milk cheese from February 19, 2014 to November 3, 2015 under this assignment. Of the total, 473 samples (29 percent) were taken from domestically produced raw milk cheese, and 1,133 samples (71 percent) were taken from raw milk cheese of international origin.
The FDA sought to design its sample collection to approximate the ratio of domestically made versus imported product on the U.S. market but was unable to do so in this case given the absence of market-share data on raw milk cheese. Thus, the agency based its sample collection ratio largely on the fact that many popular raw milk cheeses, such as Gouda, Provolone and Roquefort are traditionally of international (e.g., European) origin. The FDA collected samples to ensure that they were representative of the lot and to help enable the agency to obtain cross sections of manufacturer types for the domestic samples and countries of origin for the import samples.
Domestic Collection As directed by the assignment, FDA field staff collected 473 domestic samples of raw milk cheese from three types of establishments: manufacturers, distribution centers or warehouses,