Since they were first described (1), Escherichia coli O157: H7 and other related enterohaemorrhagic E. coli (EHEC) have become known as a major infectious cause of bloody diarrhoea. These E. coli produce one or more shiga-toxins (stx) or Vero cytotoxins. Strictly, the term EHEC refers only to those serotypes that cause a clinical illness similar to the one caused by E. coli O157:H7, but there is no widely agreed definition of when a stx-producing E. coli is considered to be an EHEC. Between 5% and 10% of infected people, particularly young children and elderly people, develop a severe complication, the haemolytic uraemic syndrome (HUS) (2,3). HUS is characterized by anaemia, a low platelet count, and renal failure, with a case fatality rate of between 2% and 7% (4) and a rate of long-term sequelae, such as renal impairment, neurological injury, or hypertension, in 12% to 30% of the cases (2,5). The inability of E. coli O157:H7 to ferment sorbitol is used widely to differentiate it bacteriologically from other E. coli (2,4), but strains of E. coli O157 that can ferment sorbitol have since been identified. The main reservoir for EHEC is cattle and other ruminants, and many outbreaks have been associated with beef products and raw milk (2). A wide range of other food products have been implicated in outbreaks of EHEC infections, such as cheese, yoghurt, fermented sausage, apple juice, seed sprouts, and lettuce (2,4,7,8). Contaminated water and direct or indirect contact with animals are other routes of transmission (2,4). The occurrence of person to person spread in outbreaks provides evidence that the infectious dose is very low (2,4). More than 5700 cases arose in an EHEC outbreak in Japan in 1996 (4), and 20 deaths were associated with an EHEC outbreak in Scotland (9), showing the potential extent and severity of EHEC infection.

Effective surveillance of EHEC in humans is essential in order to protect the public health. Surveillance is defined as the continuous systematic collection, analysis, and interpretation of data, and dissemination of the results to those in a position to take necessary actions. To prevent outbreaks that could arise through contamination of commercial food products that are widely distributed EHEC infections need to be monitored continuously in order to detect outbreaks early and, by investigating them carefully, identify risk factors. Prevention is particularly important because antibiotic treatment of EHEC infections is not recommended on the grounds that it may cause release of toxins predisposing to HUS (2). Surveillance is also critical for monitoring the emergence of new strains, particularly with regard to their ability to ferment sorbitol, which has important implications for routine diagnostic protocols. EHEC surveillance can be done by the use of EHEC infections and/or HUS surveillance, both have advantages and disadvantages. The surveillance of EHEC infections should detect recent infections and help to detect outbreaks promptly, but making the diagnosis depends on the alertness of physicians to look for EHEC and reporting is generally poorer for diseases that are usually treated in outpatients clinics and the community. In addition, methods for diagnosis of EHEC are still developing and sophisticated techniques are needed to detect non-O157 EHEC. Surveillance of HUS is less timely, information is gathered from only the fraction of those with EHEC infections who develop HUS, and the serotype often cannot be determined since most patients stop excreting EHEC by the time HUS develops. Since patients are usually in hospital, however, HUS data tend to be easier to collect and more complete.

This report presents the results of a study whose aims were to describe the surveillance for EHEC infections and HUS in Europe, paying particular attention to the general methods and the case definitions used and to compare the results obtained by these systems in terms of the number of reported cases, the presence of different serotypes, and outbreaks detected.

Methods

A questionnaire was sent to the national public health institutes in 15 European countries in March 1997. The questionnaire asked for information about the surveillance of EHEC infections, the surveillance of HUS, the recommended laboratory methods for diagnosing EHEC infections, and details of recently investigated outbreaks. All countries responded.

For the purpose of this study, we defined three categories of data source: statutory notification of disease, sentinel reporting by selected laboratories, hospitals, or practices, and varying sources from which data about EHEC infections or HUS cases were collected sporadically. The first two categories were considered to be surveillance systems because they were based on continuous systematic collection of data that allowed the potential analysis of trends.

Results

EHEC infections

Seven countries have surveillance systems (table 1). EHEC infection is statutorily notifiable in three countries - Austria, Finland, and Sweden. Five countries - Belgium, Finland, Italy, Netherlands, and the United Kingdom (UK), have sentinel systems. England, Wales, and Scotland have comprehensive national laboratory reporting schemes for Vero cytotoxigenic E. coli (VTEC) O157. Ad hoc sources exist in nine countries. Some countries have more than one data source to collect information about EHEC infections and two countries reported collecting no data on EHEC infections.

Tableau 1 / Table 1 : Sources des données sur les infections à EHEC et les SHU dans 15 pays européens (année d'introduction) / Sources for data about EHEC infections and HUS in 15 European countries (year of introduction)

Case definitions for the surveillance of EHEC infections

Four of the seven countries with surveillance systems use case definitions for EHEC infection. In Finland, the case definition is " isolation of EHEC " without further specification. Sweden and the UK use " laboratory confirmed VTEC O157 ". The case definition in the Netherlands is isolation of E. coli O157 for which the production of verotoxin and the presence of a gene common to all E. coli O157 (eaeA) has been shown. Belgium has no formal case definition, but the number of isolates of verotoxin producing E. coli strains are reported.

HUS cases

HUS is not statutorily notifiable in Europe at present. Austria, France, and Switzerland have established sentinel surveillance systems in the past three years. The UK and the Republic of Ireland began a comprehensive active surveillance system for HUS in 1997. In Italy, a sentinel surveillance system was established in 1988 (table 1). Six countries obtain information about HUS cases from ad hoc sources, and three countries reported that they do not collect data on HUS.

Case definitions for the surveillance of HUS

Five of the six countries with sentinel surveillance systems use case definitions based on triad of renal failure, anaemia, and low platelet count. Three countries set an age limit (< 15 years in France, <16 years in Ireland and the UK). The only variable uniformly defined is anaemia (haemoglobin <10 g/l plus fragmented red blood cells); criteria for the other parameters vary.

Recommendations for routine bacteriological diagnosis

Three countries have recommendations for routine bacteriological diagnosis of EHEC infections, and two countries (Germany and Italy) are preparing such recommendations. Finland recommends culture on Sorbitol-MacConkey-Agar (SMAC), and the UK recommends SMAC with cefixime and tellurite (CT-SMAC). Both media detect sorbitol-negative O157:H7, but not other serotypes of E. coli. In Denmark, DNA-probes on suspect colonies from enteric media are recommended.

Results obtained by these systems

Number of reported EHEC infections in Europe in 1996

Ten countries reported EHEC infections in 1996. The incidence in 1996 varied between 0.1 cases per million inhabitants in Spain to 20.3/1 000 000 in the UK (table 2).

Tableau 2 / Table 2 : Nombre de cas rapportés d'infections à EHEC en Europe, 1996 / Number of reported EHEC infections in Europe, 1996

Dans les pays suivants, les infections à EHEC comprennent O157 et non-O157 : Italie (5 non-O157), Danemark (3 non-O157), Allemagne (62 non-O157), Belgique (31 non-O157). Dans tous les autres pays, les nombres concernent uniquement les O157.
In the following countries, EHEC infections include O157 and non-O157: Italy (5 non-O157), Denmark (3 non-O157), Germany (62 non-O157), Belgium (31 non-O157). In all other countries, the numbers refer to O157 only.

Trends of EHEC infection

Table 3 presents time trend data from 1992 to 1996 for the four countries that reported more than 20 EHEC infections in 1996. Sentinel surveillance in Belgium was introduced in 1994, when 29 infections were reported, followed by 38 in 1995 and 52 in 1996. The increase in 1996 was due to the analysis of faecal specimens from patients with HUS by the reference laboratory in the frame of a national multicentre study. Data sources in Germany vary from year to year. The numbers of EHEC infections show a marked increase between 1993 and 1994/95 from 32 to 195 reported infections, and between 1994/95 and 1996 from 195 to 314 reported infections. In Sweden, five EHEC infections were reported from 1992 to 1994. The number rose to 114 in 1995 after a large outbreak. Subsequently, EHEC infections became notifiable in 1996, in which year 118 infections were reported. In the UK, about 600 infections were reported each year from 1992 to 1994. This increased from 685 in 1994 to 1138 in 1995 and 1180 infections in 1996 (which included the outbreak in central Scotland).

Tableau 3 / Table 3 : Tendance des infections à EHEC dans des pays européens 1992-1996* / Trend of EHEC infections for European countries 1992-1996*

* Seulement les pays qui ont rapporté plus de 20 infections à EHEC en 1996 / Only countries which reported more than 20 EHEC infections in 1996
** Les chiffres de 1994 et 1995 sont regroupés / The figures for 1994 and 1995 are combined
n.a. : non disponible / not available

Reported outbreaks

From 1992 to 1996, seven countries reported 67 outbreaks caused by EHEC. Fifty-six were reported by the UK (39 from England and Wales, 17 from Scotland) and 11 by the rest of Europe.

We received details about 47 outbreaks (table 4). Food was the likely vehicle of transmission in 23 of the 47 (including four that were spread both by food and from person to person, and one that was both food- and waterborne), seven transmitted from person to person, three were due to animal contact, one was waterborne, and in 13 outbreaks the mode of transmission remained unknown. Regarding the serotypes involved, the sorbitol-negative E. coli O157 was identified in 42 of these 47 outbreaks. But in six of them, other serotypes were identified, including one outbreak in Italy where three different serotypes (O157, O111, O86) were identified.

Tableau 4 / Table 4 : Description des 46 épidémies rapportées d'infections à EHEC ou de SHU en Europe, 1992-1996 / Details for 46 reported outbreaks of EHEC infections or HUS in Europe, 1992-1996

* sf = souche de E. coli O157 fermentant le sorbitol / sorbitol-fermenting strain of E. coli O157
n.a. = non disponible / not available

Pour l'Angleterre et le Pays de Galles, les données de 1992 - 1994 sont tirées de la réf. 14, pour 1995 et 1996, les données ne sont pas encore publiées / For England and Wales, the data for 1992 - 1994 are taken from ref. 14, for 1995 and 1996, the data are not yet published

Discussion

Currently, only half of the countries in the European Union have established surveillance systems for EHEC infection. The other countries rely on other data sources, which makes it difficult to obtain and interpret trend data. The 200-fold difference in the rate at which European countries report EHEC infections and the fact that nearly all of the outbreaks have been reported by one country, suggest markedly varying sensitivity of the data sources rather than real differences, although the possibility of real differences cannot be excluded. If all countries adopted surveillance systems, data from different countries could be compared and the value for determining trends would increase.

Case definitions are needed to ensure consistency between data from different sources and collected at different times, but only four countries use case definitions for EHEC infection. These definitions vary in their specificity and in terms of which strains are reportable. Since different techniques can be used in the laboratory diagnosis of EHEC infections, a standardised approach with recommended procedures could increase the consistency and comparability of results among and within countries (4,10,11). So far, recommendations for routine diagnostic procedures exist in three European countries.

Three of the four case definitions for EHEC used in European countries exclude non-O157 EHEC. In the fourth case, the case definition does not exclude non-O157 EHEC, but the methods recommended cannot detect them. Moreover, the methods for diagnosing non-O157 EHEC are available only in specialised laboratories. This means that information about non-O157 is very limited, which makes it difficult to assess their importance. The reported outbreaks due to non-O157 EHEC show that these serotypes occur not just in sporadic cases, however and non-O157 EHEC should be considered whenever laboratory recommendations and case definitions are developed.

Surveillance of HUS has been established in the past three years in five of the six countries where it occurs. Comparable trend data are not yet available, but the case definitions used are quite comparable. The diagnosis of HUS does not depend on stool culture and data about HUS are easier to collect because the patients tend to be in hospital, therefore the incidence of HUS may be used as a consistent marker of EHEC infections over time (4). If it can be assumed that the proportion of people infected with EHEC who develop HUS (5% to 10%) is relatively constant, data on the number of cases of HUS can be used to estimate the true number of EHEC infections evaluated. Thus, the sensitivity and coverage of EHEC surveillance can be evaluated independently. We argue, therefore, that HUS surveillance should be adopted as well as EHEC surveillance.

Currently approaches to the surveillance of EHEC related disease and the laboratory methods used to diagnose it differ from one country to another. Data within Europe are not comparable at present. In February 1997, an outbreak of E. coli O157 infection in the Canary Islands affected people from several European countries (12). In order to prepare for future international events, 'Enter-net', a European Union concerted action under the BIOMED 2 programme, is planning to extend its activities on EHEC. It is developing a specific monitoring system to determine the distribution of EHEC, and establishing a core set of data items to accompany each laboratory typed EHEC isolate (13). These data will create an international database readily available to each participating team in order to detect clusters of EHEC isolates in time, place, and person, and to bring such clusters quickly to the attention of collaborators.

Particular thanks go to the members of the national public health institutes, national public health laboratories, and the EPIET fellows who have provided the data reported in this article.

Collaborateurs / Collaborators:

EPIET-Fellows and the following staff members of the Public Health Institutes and Laboratories:

F. Allerberger, Bundesstaatliche bakteriologisch-serologische Untersuchungsanstalt, Austria;
D. Pierard, Academisch Ziekenhuis Vrije Universiteit Brussels, Belgium;
P. Gerner-Smidt, Statens Seruminstitut, Denmark;
G. Adak, CDSC London, United Kingdom;
A. Siitonen, National Public Health Institute, Finland;
B. Decludt, Réseau National de Santé Publique, France;
A. Karaitianou-Velonaki, Ministry of Health and Welfare, Greece
J. Kiely, Department of Health, Ireland;
A. Tozzi, Istituto Superiore di Sanità, ltaly;
M. de Wit, Rijksinstituut voor Volksgezondheid en Milieuhygiene, The Netherlands;
E. Mitchell, DHSS, Northern Ireland, United Kingdom;
M.T. Paixao, Instituto Nacional de Saúde, Portugal;
W. Reilly, Scottish Centre for Infection and Environmental Health, Glasgow, United Kingdom;
J.F. Martinez-Navarro, Instituto de Salud Carlos III, Spain;
B. de Jong, Swedish Institute of Infectious Disease Control, Sweden;
H. Schmid, Bundesamt für Gesundheit, Switzerland.

1. Riley LW, Remis RS, Helgerson SD, McGee HB, Wells JG, Davis BR, et al. Haemorrhagic colitis associated with a rare Escherichia coli serotype. N Engl J Med 1983; 308: 681-5.

2. Griffin PM. Escherichia coli O157:H7 and other enterohaemorrhagic Escherichia coli. In Blaser MJ, Smith PD, Ravdin JI, Greenberg HB, Guerrant RL. Infections of the Gastrointestinal Tract. New York: Raven Press Ltd, 1995.

3. Seibt F, Filler G, Gellermann J, Beutin L, Ehrich JHH. The heterogeneity of haemolytic uraemic syndromes in children and adults. Nephrologisches Jahresgespräch 1995; Editor: Deutsche Dialysegesellschaft.

4. WHO. Prevention and control of enterohaemorrhagic Escherichia coli (EHEC) infections. Geneva: WHO, 1997.

5. Bockemühl J, Karch H. Assessment of the actual importance of enterohaemorrhagic Escherichia coli (EHEC) in Germany (1994-1995). Bundesgesundhbl 1996; 39: 290-6.

6. Tarr Pl. Escherichia coli O157:H7: clinical, diagnostic, and epidemiological aspects of human infection. Clin Infect Dis 1995; 20:1-8.

7. Outbreaks of Escherichia coli O157:H7 infection and cryptosporidiosis associated with drinking unpasteurized apple cider - Connecticut and New York, October 1996. MMWR Morb Mortal Wkly Rep 1997; 46:5-8.

8. Hilborn ED, Mermin J, Mshar P, Slutsker P, Mshar R, Farrar J et al. A multistate outbreak of Escherichia coli O157:H7 infections associated with mesclun mix lettuce. Abstract in: Program and abstracts of the 46th annual epidemic intelligence service (EIS) conference. Atlanta: Centers for Disease Control and Prevention, Atlanta, 1997.

9. Anonymous. Late death arising from the central Scotland outbreak of infection with Escherichia coli O157. SCIEH Weekly Report 1997; 31:113.

10. Pennington TH. Report on the circumstances leading to the 1996 outbreak of infection with E. coli O157 in Central Scotland, the implications for food safety and the lessons to be learned. Edinburgh: The Stationery Office, 1997.

11. Alexander ER. Editorial response: surveillance of Escherichia coli O157:H7 - a necessity for the prevention of an emerging disease? Clin Infect Dis 1994; 19:814-5.

12. CDSC. European collaboration identifies an outbreak of Escherichia coli O157 infection in visitors to Fuerteventura, Canary Islands. Comm Dis Rep CDR Wkly 1997, 15:127.

13. Fisher I on behalf of the Salm-Net collaboration. Enter-net: International laboratory based surveillance of salmonella and VTEC O157 infection. In: Abstracts of the international symposium: European study group on enterohaemorrhagic Escherichia coli, Innsbruck, April, 1997. Fortschritte der Medizin, 1997; Monographie 84:56.

14. Wall PG, McDonell RJ, Adak GK, Cheasty T, Smith HR, Rowe B. General outbreaks of Vero cytotoxin producing Escherichia coli O157 in England and Wales from 1992 to 1994. Commun Dis Rep CDR Rev 1996; 6:R26-33.