A print friendly PDF version of the full article is available from this Communicable Diseases Intelligence issue's table of contents.
The OzFoodNet Working Group
This fourth annual report of OzFoodNet highlights the burden that foodborne illness places on the health system and community. The cost to Australia each year from foodborne disease may be as high as $AUD1.2 billion annually.18 In recent years, Australia has experienced consistently increasing rates of notified infections, along with increasing numbers of foodborne outbreaks. In 2004, we observed a 28 per cent increase in reported foodborne outbreaks compared with 2003, which may be due in part to improved surveillance. However, there has not been any appreciable change in surveillance system for notifiable diseases. This is a concern, as there was a nine per cent increase in notifications of potentially foodborne diseases in this report when compared to historical averages.
The United States of America program—FoodNet—recently reported significant declines for 2004 in the incidence of human salmonellosis and campylobacteriosis, which were attributed to improvements in agricultural industry and reduction in isolation of Salmonella and Campylobacter in food processing plants.9 Australia could consider gathering similar data on Salmonella and Campylobacter in animals and developing disease reduction targets to focus prevention efforts.
The major causes of foodborne disease in Australia during 2004 were similar to previous years, with fish, poultry, bakery products, seafood, and eggs being the major causes of outbreaks. While seafood and fish are responsible for large numbers of outbreaks they are usually small in size and are rarely associated with Salmonella and Campylobacter infections, which make up the majority of sporadic infections reported to health departments.19 In contrast, poultry and eggs, are common causes of these sporadic infections. Risk of campylobacteriosis is strongly associated with consumption of under-cooked chicken in Australia and may be responsible for between 5–11 per cent of infections.20 OzFoodNet sites reported several outbreaks associated with chicken during 2004.
The largest of these chicken-associated outbreaks was a community-wide outbreak of Salmonella Typhimurium 12 in New South Wales that affected 141 people. These community-wide increases occur commonly and are difficult to investigate due to poor patient recall of foods consumed and the high frequency of chicken consumption. Chicken meat is commonly contaminated with Salmonella and Campylobacter at retail sale.21,22 While cooking readily kills these bacteria, reducing the concentration of bacteria on meat through control measures on farms or in processing plants could possibly lower the incidence of these diseases in the community. The S. Typhimurium 12 outbreak highlighted the potential of chicken as a vehicle of community outbreaks and the need to reduce contamination of raw meats through improved primary production and processing.
There were many outbreaks in 2004 where investigators were unable to confirm the aetiology. One reason for this is that surveillance of outbreaks is improving with smaller outbreaks being detected where it is difficult to confirm an aetiological agent. Another reason is the changing nature of laboratory tests. The availability of test kits for C. perfringens in Australia was limited in 2004, due to concerns about importation of Clostridium toxins. This made it difficult to confirm Clostridium perfringens as the cause of outbreaks, as traditional case definitions rely on confirming greater than 105 organisms in two or more faecal specimens or the demonstration of C. perfringens toxin.23
Similarly, epidemiologists were unable to identify a food vehicle in 63 per cent of outbreaks investigated using analytical epidemiological studies. In many of these outbreak investigations epidemiologists suspected a vehicle, but did not have the epidemiological, microbiological or traceback evidence to implicate a specific food. It can be difficult to effect and implement food recalls in these instances.
The outbreaks of norovirus associated with Individually Quick Frozen oyster meat were an example of where epidemiological and traceback evidence confirmed a food vehicle, but virological analysis was negative for norovirus. National discussions with food safety agencies identified the risks associated with these products, and Importers agreed to withdraw products from the marketplace. During 2004, the Northern Territory Health Services submitted oyster meat from an outbreak in the previous year to Environmental Science and Research in New Zealand, which confirmed contamination with norovirus (personal communication, Gail Greening, 3 May 2005). This positive test result, months after the initial investigation, vindicated various food safety agencies decisions to seek the withdrawal of these products. The small outbreak in Queensland in October several months after the original withdrawal showed how long-shelf life products can cause problems for public health agencies conducting food recalls. As a preventive measure, the Australian Quarantine and Inspection Service has implemented restrictions to imports of oyster meat from the specific harvest area where implicated oysters were harvested, and certain Korean oyster growing areas that have supplied products causing outbreaks in New Zealand.
Norovirus caused a considerable amount of gastroenteritis that was not foodborne in 2004. OzFoodNet sites reported a massive number of outbreaks spread from person-to-person, many of which occurred in aged care and hospital settings. Norovirus was responsible for 45 per cent of outbreaks spread from person-to-person and accounted for 13,739 cases. The genotype of these norovirus outbreaks was reported for very few of these outbreaks, as many were diagnosed using rapid enzyme-based detection test kits. It is likely that many of these outbreaks would have been due to the new variant of genotype II4 that affected Australia during 2004 and other countries internationally.24 Epidemic clones of norovirus can affect multiple countries and cause widespread illness in community and institutional settings.25 It is also likely that many of the person-to-person outbreaks of unknown aetiology would have been due to norovirus. The high burden of disease in the healthcare system is very costly and it is important for public health agencies to be able to identify interventions that are able to halt the spread of these outbreaks.26
OzFoodNet identified several risk factors for foodborne infections in 2004 based on the surveillance data and epidemiological studies. Many of these risk factors have been previously recognised, but may need to be considered again. Some of the risks require complex solutions, while others are far more simple, such as using pasteurised eggs in food service industries or avoiding consumption of shellfish and fish harvested from high-risk areas. While Australia does not have endemic S. Enteritidis 4 that contaminates the internal contents of eggs, there are clearly other subtypes of Salmonella that are associated with eggs. As noted in previous years, there is a need to identify the potential food safety failures in bakeries to prevent outbreaks of salmonellosis.27
It is important to recognise some of the many limitations of the data that OzFoodNet reports. Surveillance data are inherently biased and require careful interpretation. These biases include the higher likelihood that certain population groups will be tested, and different testing regimes in different states and territories, resulting in different rates of disease. In some jurisdictions, the rates of disease are unstable due to small numbers of notifications and populations under surveillance. Importantly, some of the most common enteric pathogens are not notifiable, particularly norovirus, Clostridium perfringens and enteropathogenic E. coli. These organisms may be notified as the cause of outbreaks, but not as individual cases of disease. There can also be considerable variation in assigning causes to outbreaks depending on investigators and circumstances. States and territories are moving towards harmonising surveillance as much as possible to address some of these issues.
Surveillance of foodborne diseases in Australia has steadily improved in recent years, as shown by data on improving completeness of Salmonella subtyping on state and territory surveillance databases. OzFoodNet aims to continually improve surveillance and investigation practices regarding enteric and foodborne diseases. In 2004, the capacity of Australia to respond to multi-state foodborne disease outbreaks was examined in an external review. The review found that in recent years there had been significant improvements in capacity, but there was still room for improvement. Following the consultation, OzFoodNet conducted a trial of a web-based database for capturing individual patient data during multi-state outbreaks.28 The trial was based on a mock outbreak of a fictitious Salmonella serotype—Mordor—amongst hobbits, wizards, ents and men after a hobbit's birthday party. The trial showed the benefits of using web-based databases for multi-site investigation of outbreaks. The issues highlighted in the review are common to many countries that investigate foodborne illness across multiple jurisdictions.29
The large burden of foodborne disease observed in 2004 is a concern for Australia. In future years, it may be appropriate to set targets for reducing foodborne disease in conjunction with other government agencies and industries. It is important that foodborne disease surveillance is able to assess whether food safety policies and campaigns are working. National surveillance of foodborne diseases has many benefits and provides long-term data to review causes of illness. Since OzFoodNet began surveillance in 2000, the network has collected information on the cause of approximately 400 outbreaks of foodborne disease. These data are becoming useful for reviewing less common, but important, causes of disease outbreaks. Ongoing efforts are needed to strengthen the robustness of these data and ensure that they continue to be useful to agencies developing food safety policy.
We would like to thank the many epidemiologists, project officers, interviewers and research assistants at each of the Sites who contributed to this report. We also acknowledge the work of various public health professionals and laboratory staff around Australia who interviewed patients, tested specimens, typed isolates and investigated outbreaks. The high quality of their work is the foundation of this report. OzFoodNet is an initiative of the Australian Government.
This article was published in Communicable Diseases Intelligence Vol 29 No 2, June 2005.