The average annual notified rate of STEC illness over the past 11 years (2000-2010) in Australia using NNDSS data was estimated to be 0.4 cases per 100,000 per year, with an average rate of confirmed O157 STEC illness of 0.12 cases per 100,000 per year. Although it is difficult to see a clear trend from these national data, rates look to have increased slightly since 2000. Whilst this may represent a real increase in the STEC illness rate, there is also evidence that the increase may be linked to an increase in the number of stool samples tested and changes in laboratory methods used to detect STEC (unpublished observations).
Variation in rates across states may also be partly related to variation in testing practices. The number of stools tested for STEC varies considerably between different states of Australia and this appears to be correlated with STEC notifications rates. The laboratory tests used to detect STEC also varies between jurisdictions, with some using sorbitol MacConkey agar to culture STEC and others incorporating PCR for detection of stx1
genes, which is considered the more sensitive method [37
]. The Australian state that tests the most number of stools is South Australia even though it does not have a large population compared with many of the other states of Australia. In this jurisdiction, a single reference laboratory tests specimens that include those with evidence of blood in the stool, where there is a clinical history of blood in the stool, or where the requesting doctor queries STEC infection or HUS. The specimens tested are referred from other private and public health laboratories and so this jurisdiction is considered to provide Australia with amongst the most robust estimates internationally for STEC incidence in bloody diarrhoea [24
]. However, despite the quality of the surveillance in this state it is not entirely clear if STEC rates and epidemiology can be extrapolated from South Australia to other regions of the country, as Australia is also a very large country with differing geography, climate and demographics.
South Australian surveillance data showed that the 11 year average rate for STEC infections was 2.4 cases per 100,000 per year, which is over 6-fold greater than the national rate over this period. South Australian STEC incidence rates have remained fairly steady over a considerable period, apart from the increase in 2009 due to several outbreaks in this jurisdiction in that year. As there have been no major changes in surveillance practices in South Australia over the last 15 years, this supports the notion that there has been no real change in the rate of STEC infections in Australia over this period.
Surveillance practices also vary considerably between countries and therefore caution is required when comparing STEC incidence rates between countries. In particular, laboratory testing practices vary considerably, especially the practices for screening stools for blood and the extent to which genetic testing methods are utilised, and in addition, the extent to which testing for all serotypes occurs. In Australia, testing for non O157 serotypes is an important component of surveillance with most jurisdictions conducting STEC testing also using methods to detect non O157 serotypes [24
]. However, testing for non O157 serotypes does not occur in many countries, with most focussing mainly on the detection of O157 strains [24
]. In addition, case definitions are not the same across countries and some reports are for 'confirmed' cases only, while others include 'probable' cases. Some countries have passive or voluntary reporting, while others are mandated by law or actively pursued. Furthermore, multiple STEC infections in one person may be reported as separate infections or as one case. Even within countries there is variation in testing and reporting. For example, in the Netherlands only some parts of the country test for non-O157 [38
]. Also, in the United States, only 7% of 428 clinical laboratories surveyed used enzyme immunoassay, a non-culture method for the detection of non O157 cases [40
Bearing in mind the need for caution regarding interpretation, Australian rates appear to be lower than countries in Europe that have similar surveillance practices and report confirmed cases of both O157 and non O157 STEC. Even the higher rates obtained from the South Australian surveillance system are comparable to rates reported in these countries. The Community Summary Report on the European Union in 2008 [38
] gives the incidence of all STEC infections in 2008 in Austria, Belgium and Norway as 1 case per 100,000 or less; and Denmark, Sweden and Ireland as 2.9, 3.3 and 4.8 cases per 100,000, respectively. New Zealand, Australia's nearest neighbour, reported STEC in 2009 at a rate of 3.3 cases per 100,000 per year [42
]. In many countries the vast majority of reported STEC cases were serotype O157, including Scotland, Ireland, the United Kingdom as a whole, the United States and Canada. In Scotland in the period 1998-2007, the annual average incidence of O157 infections was 4.3 cases per 100,000 per year [39
]. In Ireland in 2009, O157 incidence was 3.9 cases per 100,000 per year [43
] whilst in the whole of the United Kingdom in 2008 it was reported to be 1.9 cases per 100,000 per year [38
]. In the United States, between 2000- 2006, data gathered from eight FoodNet surveillance sites gave an incidence of O157 of 1.5 cases per 100,000 per year [44
]. In Canada in 2007 the incidence of O157 was 2.9 cases per 100,000 per year [41
]. It should be noted that all of these rates are higher than the incidence rate for confirmed O157 for South Australian of 0.7 cases per 100,000 per year in the period 1997-2009.
The overall incidence of HUS in Australia in the 11 year period 2000-2010 was 0.07 per 100,000 per year with children under 5 years having the highest rate of 0.49 cases per 100,000 per year. It is important to recognise that HUS can be due to various causes and not all cases are secondary to STEC infection. An estimate by a group of Australian foodborne disease experts in 2005 suggested that about 50% of cases were secondary to STEC infection [45
]. Other evidence suggests the proportion could be as high as 88%, with 139 out of 160 HUS cases reported to the Australian Paediatric Surveillance Unit between 1994 and 2001 being diarrhoea associated [46
National hospital separation data showed more episodes of hospitalisation for HUS than the number of cases notified to national surveillance. Firstly, it is likely that patients with HUS may have multiple hospital admissions, especially as they may be transferred to specialist units and so are recorded as more than one hospital admission episode. Furthermore, any cases proceeding to dialysis may require many admissions. Cases from earlier years that developed chronic renal failure due to HUS may still appear in this hospitalisation data in this time period for treatment. It is also possible that some hospitalised cases are not reported to the national surveillance system. There may be some cases that were not STEC but other types of E.coli infection included in the codes used to identify episodes of hospitalisations attributed to STEC. On the other hand, there are also likely to be other STEC infections coded to diarrhoea of presumed infectious origin. A validation study linking notifications, hospitalisations and deaths would be valuable to clarify this issue.
Worldwide, case control studies of varying sizes and rigour have been conducted aiming to examine risk factors for sporadic O157 STEC infection. In many of the larger well conducted North American studies a significant association between illness and the consumption of hamburgers, pink or undercooked hamburgers, pinkish ground beef, undercooked meat, or barbequed food, has been demonstrated [47
]. In addition to this, eating in restaurants or fast food restaurants were also identified as risk factors [48
]. Living or working or visiting a cattle farm [47
] and visiting a farm with cows [48
] were also both found to be strongly associated with O157 STEC infection in two of the larger United States studies completed. Drinking untreated surface water, drinking well water, or swimming in a pond were other environmental exposures associated with infection in these North American studies. Other case control studies have indicated the association between illness and contact with animal faeces [52
], consumption of raw milk [53
] and consumption of cold cooked sliced meats [54
]. Evidence of household transmission has also been obtained, with an association between STEC infection and the presence in the household of a child under the age of 2 years [48
], a child with diarrhoea [55
], or anyone with diarrhoea [49
In Australia, two case control studies have examined risk factors for sporadic STEC infection. In South Australia in 2002 cases were more likely than controls to have eaten berries, including strawberries, blueberries and blackberries, in the 10 days preceding illness [56
]. This finding is interesting, although the study was small and this finding needs to be interpreted with caution. Another Australian study [57
] recruited 114 cases and risk factors for infection were analysed for those infected with O157 serotypes and non-O157 serotypes of STEC. Risk factors for O157 infection in this study included eating hamburgers, visiting restaurants, having previously used antibiotics, having direct contact with red meat through occupational exposures or having another member of the household having occupational exposure to red meat. The risk factors for O157 infection in this Australian study were largely similar to those reported in international studies; however, consumption of hamburgers and ground beef has not been previously thought to be an important cause of STEC infections in Australia so this finding is of considerable interest. Risk factors for non O157 infection included having eaten sliced chicken meat or corned beef from a delicatessen, having camped in the bush, having eaten catered meals, or having had family occupational exposure to animals. The grouping of all non O157 serotypes in this study into one group assumes that different serovars have similar characteristics and risk factors which is a limitation; however, the identification of different risk factors for O157 and non O157 STEC almost certainly reflects differing reservoirs for these serogroups in Australia.
STEC outbreaks appear to affect less people in Australia compared to many countries. A total of 117 cases were reported as part of 11 outbreaks in Australia as a whole in the nine year period between 2001-2009, a crude rate of 0.5 outbreaks per million population per year and representing approximately 15% of all STEC cases reported. This is much lower than the proportion of outbreak cases in some other comparable countries. For example, in Ireland, 50% of cases in 2009 were related to 42 outbreaks in that year [43
]. In New Zealand between 2006 and 2009 there were 19 outbreaks of STEC, three of them foodborne [42
]. In 2008 in the European Union, there were 75 foodborne outbreaks of STEC and other pathogenic E.coli
, up from 65 outbreaks in 2007 [59
]. The largest numbers of outbreaks reported from a single country were in Germany (28 outbreaks) and Austria (11 outbreaks) with only four reported from the United Kingdom. In the United States in 2007 there were 40 outbreaks of STEC [60
]. In the United States, ground beef appears has posed the most significant public health risk and has been the food most commonly implicated in reported STEC outbreaks, although over time there have been major outbreaks attributed to other sources, such raw milk products, fresh produce, juice, sprouted seeds and spinach [12
]. In 2007, five out of six multi-state outbreaks of E.coli
O157:H7 were due to ground beef [60
]. Many of the outbreaks found to be due to ground beef over the years have been extremely large in size and have affected up to many hundreds of individuals [9
]. The majority of these outbreaks were most likely to have been due to beef where contamination was disseminated in product during the grinding process and then undercooked during preparation [3
]. It is important to mention that in addition to the 11 outbreaks in Australia reported between 2001 and 2009, there was also a significant STEC outbreak in South Australia in 1995, with 23 HUS cases in children and one death associated with the consumption of mettwurst, a fermented but uncooked meat product [27
The epidemiology of STEC serotypes suggests O157 strains have been less dominant in Australia than the United States, England, Wales, Scotland and Japan, both in terms of sporadic cases, and particularly as a cause of outbreaks. As mentioned earlier, in many countries, including the United States [40
], O157 has been the focus of testing regimes, and this goes some way towards explaining the predominance of this strain in surveillance data from these countries. In Australia, testing for non O157 serotypes does occur routinely and has shown that O111, O26, O113, O55 and O86 strains are also commonly associated with STEC illnesses.
The burden of illness associated with STEC and HUS is considerable. A study exploring the economic costs of STEC infection in South Australia between 2003 and 2006 found that of the 3-7% of sporadic STEC patients who developed HUS, 40% had ongoing medical issues [63
]. In this study, 19 out of 43 STEC cases were admitted to hospital (44%). The estimated average cost of STEC infection was AUD 3,132 per case. It is also likely that in every country, the reported burden is an underestimate and that there are cases in the community that are not tested and reported [64
]. In South Australia, it has been estimated that there are around eight (95% credible interval 3-75) cases in the community for every case detected by surveillance [65
]. After accounting for under reporting of STEC to surveillance, this equated to a cost of AUD 2,633,181 for Australia each year.