Part I RisksRisks
2.1 Introduction: the risk of microbial foodborne disease
Accurate estimates of foodborne diseases are difficult to make because of underreporting. If estimates are made, however, one can see that orders of magnitude of the various organisms differ so much that even with these uncertainties, global conclusions remain appropriate. It is estimated that in the USA there is a chance of about 1 in 3 per year of getting a foodborne illness, a chance of 1 in 800 of getting hospitalised, and a chance of 1 in 55 000 of dying (Mead et al., 1999). It is useful to compare this estimate with other causes of death, to put it in perspective (Table 2.1). From this comparison one can conclude that although foodborne diseases are not among the most relevant causes of death, they are relevant and more important than minor causes such as lightning.
2.1.1 Microorganisms responsible for foodborne diseases
Comparing the various organisms can also show relevant trends (Table 2.2), we can see that in the USA noro-viruses, Campylobacter and Salmonella are responsible for more than 90% of the diseases, and for deaths Toxoplasma and Listeria are also of relevance. So only five organisms give rise to more than 90%
of the problem (Mead et al., 1999). That does not mean that one should not look at other organisms, since these are only estimates, which may change over time.
The comparison of different countries will result in different outcomes, because of real differences in risks since other products, procedures, habits are used, but also because of different surveillance systems. The contributions of
2
The range of microbial risks in food processing
M. H. Zwietering and E. D. van Asselt, Wageningen University, The Netherlands
Table 2.1 Estimated chance of dying (per year) by different causes*
P dying Number per million Cause
1:115 8800 Total
1:10 000 100 Infectious disease
1:10 000 100 Suicide
1:15 000 75 Traffic accident
1:55 000 18 Foodborne disease
1:200 000 6 Drowning
1:10 000 000 0.1 Natural disasters
1:20 000 000 0.05 Lightning
* All data (except foodborne diseases, which are from the USA; Mead et al. (1999)) are based on Dutch statistical data.
Table 2.2 Relative importance of various causes for disease, hospitalisation and death (Mead et al., 1999)
Illness (%) Hospital (%) Death (%)
Bacillus cereus 0.198 0.014 0
Botulism 0.00042 0.076 0.246
Brucella 0.0056 0.100 0.306
Campylobacter 14.2 17.3 5.7
Clostridium perfringens 1.8 0.064 0.360
Escherichia coli 1.3 4.6 4.3
Listeria monocytogenes 0.018 3.8 27.5
Salmonella non-typhoidal 9.7 25.7 30.4
Shigella 0.649 2.0 0.790
Staphylococcus 1.3 2.9 0.107
Streptococcus 0.369 0.586 0
Vibrio 0.038 0.203 1.7
Yersinia enterolitica 0.628 1.8 0.126
Cryptosporidium parvum 0.217 0.327 0.365
Cyclospora cayetanensis 0.106 0.025 0.021
Giardia lamblia 1.4 0.822 0.055
Toxoplasma gondii 0.814 4.1 20.7
Trichinella spiralis 0.00038 0.0069 0.0086
Noro-viruses 66.9 32.9 6.8
Rotavirus 0.282 0.822 0
Astrovirus 0.282 0.205 0
Hepatitis A 0.030 0.891 0.460
Total 100 100 100
Bold figures represent the most important causative agents (values larger than 5%).
32 Handbook of hygiene control in the food industry
various pathogens to the total number of foodborne illnesses and deaths are given in Figs 2.1 and 2.2 respectively, for surveys from the USA (Mead et al., 1999), the UK (Adak et al., 2002) and France (Vaillant et al., 2004).
As described before, there is underreporting in the number of documented foodborne illnesses. To account for this underreporting, various surveys use an underreporting factor to obtain a realistic estimate of the real number of illnesses based on documented cases. The USA survey is based on surveillance systems and uses underreporting factors that are either based on literature studies or are estimated based on, for example, expert opinions (Mead et al., 1999). The UK survey (Adak et al., 2002) uses underreporting factors that are based on data from laboratory reports and the incidence rate in the population in 1995 found in a study to infectious intestinal disease. This factor is then applied to data from laboratory reports in other years. Which pathogens are most important in a country, therefore, also depends on the underreporting factor used (Table 2.3).
It can be seen that in all three countries noro-virus, Campylobacter and Salmonella are very important pathogens for causing foodborne illnesses. In the UK, Clostridium perfringens and Yersinia spp. are also in the top five of the most important pathogens. However, in these cases much higher underreporting factors are used than in the USA. Foodborne deaths in the three countries are
Fig. 2.1 The contribution of various pathogens to the total number of foodborne illnesses in percentages. No data available for Toxoplasma gondii in the UK.
The range of microbial risks in food processing 33
Fig. 2.2 The contribution of various pathogens to the total number of foodborne deaths in percentages. No data available for Toxoplasma gondii in the UK and for noro-virus in
France.
Table 2.3 Ranking of pathogens according to their contribution to the overall number of foodborne illnesses or deaths together with the underreporting factor used in the UK (Adak et al., 2002), USA (Mead et al., 1999) and France (Vaillant et al., 2004) Microorganism Ranking illness Ranking death Underreporting factor
UKa USA France UKa USA France UKa USA France
Campylobacter spp. 1 2 4 3 5 4 10 38 ±
C. perfringens 2 4 5 2 7 6 364 38 ±
L. monocytogenes 7 8 8 4 2 2 2 2 ±
E. coli O157:H7 6 7 7 5 6 7 2 20 ±
Salmonella spp. 5 3 3 1 1 1 4 38 ±
Yersinia spp. 4 6 6 7 8 5 1254 38 ±
Toxoplasma gondii ? 5 2 ? 3 3 ? 15 ±
Noro-virus 3 1 1 6 4 ? 276 ? ±
aData from 2000.
? Unknown.
± No under-reporting factor used.
34 Handbook of hygiene control in the food industry
mainly caused by Salmonella and L. monocytogenes. In the UK, Cl. perfringens is the second most important pathogen causing foodborne deaths, but again the underreporting factor is much higher than in the USA.
2.1.2 Related products
Once an important organism has been identified, it is important to identify its transmission route. One can follow three approaches:
· identify foods in outbreaks (by case-control, typing, . . .);
· comparing types in cases/foods (sero-, phage-typing);
· quantitative risk assessment.
The first method is straightforward in that if a certain food/organism combination is suspected in a case or an outbreak, one can clearly establish a link between the food product and the case/outbreak by typing. Foods suspected in incidents are usually investigated by food inspection services and the WHO has collected these data for various countries in the world (Fig. 2.3) (Rocourt et al., 2003). It can be seen that the products that contribute most to the number of foodborne illnesses are meat and meat products and eggs and egg products. This is probably caused by the presence of Salmonella and Campylobacter spp., which are known to be an important cause of foodborne outbreaks (see previous
Fig. 2.3 Contribution of various food products to the number of foodborne illnesses for various countries in 1998±2001 (Rocourt et al., 2003).
The range of microbial risks in food processing 35
section). Important considerations are that sporadic cases will often go undetected, since the focus with this method is on larger incidents. Moreover, it is often difficult to gather all relevant details of an outbreak, because of the time between the moment of consumption and the investigation.
An example of the second approach can be found in a study by Hald et al.
(2002) who compared various types of salmonellae in patients with types in various food products, resulting in estimates of the contribution of these food products to the disease burden (Fig. 2.4). An example of the third approach concerns the use of quantification to determine the importance of various routes of contamination. Evers et al. (2004) determined the contribution in the exposure of Campylobacter of various routes. This type of quantification can effectively help to target interventions. The very extended FDA/FSIS (Food and Drug Administration/Food Safety Inspection Service) risk assessment (HHS/
USDA, 2003) concerning the relative risks of Listeria monocytogenes in ready-to-eat foods resulted in the identification of the most important product groups.
Deli meats were identified as largely the most important source, followed unexpectedly by pasteurised milk. In this product, the risk per serving is not very high, but owing to the large number of consumed units for the per annum risk, and thus the number of cases in a year, it is a relevant source. The low prevalence of Listeria in this product is probably caused by recontamination (HHS/USDA, 2003).
Fig. 2.4 Estimated number of sporadic and domestic cases of human salmonellosis in Denmark in 1999 attributed to different animal-food sources (Hald et al., 2002).
36 Handbook of hygiene control in the food industry