Data from a national network of paediatric nephrology departments in France suggest that the incidence of haemolytic uraemic syndrome (HUS) in 1998 was 0.7 cases per 100 000 children aged under 15 years and that cases occur sporadically. Six out of 85 cases reported in 1998 were imported, which illustrates a need for national and international early warning systems for HUS. The proportion of cases with positive serology has fallen since this surveillance system was set up in 1996. This suggests that new serogroups of Escherichia coli undetectable by current diagnostic serology kits may be responsible for some cases of HUS.

Shiga-like toxin producing Escherichia coli (STEC), also known as Vero cytotoxin producing E. coli and enterohaemorrhagic E. coli, have been responsible for several large epidemics of food poisoning with high levels of mortality in industrialised countries (1,2). Clinical manifestations of STEC infections include diarrhoea, haemorrhagic colitis, and haemolytic uraemic syndrome (HUS), which is the main cause of acute renal failure in children between 1 month and 3 years of age (3). Two surveys of patients admitted to hospital with HUS between January 1993 and March 1996, undertaken in cooperation with the Société de Néphrologie Pédiatrique (paediatric nephrology society) (4), led to the establishment of a surveillance system for the monitoring of HUS in children, based on a national network of paediatric nephrology departments in France in 1996, coordinated by the Institut de Veille Sanitaire (InVs, the French public health institute). The objectives of this surveillance are to monitor trends in time and space of HUS in children aged under 15 years in France, determine the characteristics of cases, identify the agents responsible, and detect outbreaks.

The surveillance system defines a case of HUS as any person aged under 15 years in whom a clinical diagnosis of HUS (sudden onset of haemolytic anaemia with renal failure) has been made by a nephrologist or paediatrician on the basis of microangiopathic haemolytic anaemia (haemoglobin <10 g/100 mL and schizocytosis ³ 2%) and renal failure (plasma creatinine >60 µmol/L if age <2 years (normal £ 50 µmol/L), >70 µmol/L if age ³ 2 years (normal £ 60 µmol/L)).

A case of STEC infection is defined as any child under 15 years of age from whose faeces STEC has been isolated, genes encoding Shiga toxins (5) have been detected by polymerase chain reaction (PCR), or in whose serum antibodies against one of the 26 serogroups of E. coli tested had been found.

The surveillance of paediatric HUS in France is based on a network of 31 departments of paediatric nephrology in university and general hospitals, throughout metropolitan France (figure 1). This network has been stable since its creation. Whether notifications are comprehensive has not been evaluated, but the rate is probably very good because of the rarity and severity of the illness, the voluntary participation of paediatricians, and regular monitoring of reports by serological tests performed at the enterobacteria unit of the Institut Pasteur, Paris.

The following information was collected in the preliminary study and continues to be collected by the surveillance system: sociodemographic data, existence of prodromal diarrhoea, laboratory findings on admission, stool culture results for organisms sought routinely, deaths, participation in community life and canteen meals, occurrence of other cases of HUS in the community (family, nursery, school, village) in the previous 12 months, cases of diarrhoea in the environment during the 30 days before or after the onset of HUS, and exposure to risk factors during the two weeks before onset of first symptoms (consumption of unpasteurised milk or cheese, minced beef and method of cooking, contact with farm animals, swimming).

The current surveillance protocol requests that two serum specimens for each case, obtained on the first and 15th days of illness, should be sent to the enterobacteria unit of the Institut Pasteur, to be tested for antibodies against the lipopolysaccharide (LPS) of 26 serogroups of E. coli. Each serum is tested using two different techniques (an enzyme linked immunoabsorbent assay (ELISA) test and a ‘line blot’ test).

In 1998, 85 cases of HUS were reported, 79 of whom acquired infection in France. Six cases who became ill during or immediately after a stay abroad were considered to be imported and were analysed separately from the native cases.

In 1998, the incidence of HUS in France was 0.7 cases per 100 000 children aged under 15 years, the lowest rate since 1993. The number of cases in 1993 may have been smaller because data were collected retrospectively (table). Since the surveillance system was set up, 52% of cases (267/515)) have occurred during the summer months (June-September), but this seasonal trend was less marked in 1998, when two seasonal peaks were in spring (April-May) and summer (July-September).

Table.  Number of cases and annual incidence rate of HUS in children under 15 years of age. Metropolitan France, 1993-1998  

Year

Number of cases

Incidence / 10⁵ children < 15 years / year

1993

58

0.51

1994

94

0.82

1995

94

0.82

1996

90

0.79

1997

100

0.89

1998

79

0.70

1993-1998

515

0.76

The age distribution of cases in 1998 was similar to that in previous years, with the highest incidence in children under two years of age (3/100 000). The mean annual incidence from 1993 to 1998 has remained stable, at about 1.9/100 000 children aged under 5 years and 0.2/100 000 children aged 5 to 14 years. The highest mean incidence is among 1 year old children (3.3/100 000) (figure 2). The sex ratio of cases was 1:1 in 1998 compared to 1.1 (52 female versus 49 male) in 1993-1997.

Clinical description

In 1998, prodromal diarrhoea preceded HUS in 91% (72/79) of cases. Diarrhoea was bloody in three fifths (40) of the 66 patients for whom such information was available.

This prodromal diarrhoea led to 39 of these 66 patients being admitted to hospital. The diagnosis of HUS was made between 0 and 44 days after the onset of diarrhoea (median 7 days). From 1993 to 1998, five of the 515 cases died, two in 1993, two in 1994, and one in 1996.

In 1998, 60 (76%) of the 79 patients were tested serologically (91% (118/130) in the 1995-6 prospective study, 94% (84/90) in 1996, and 79% (79/100) in 1997). Antibodies against one or more of the 26 serogroups tested were found in 29 of these 60 patients. Serogroup O157 (alone or in association) was found in 27 cases (93% of cases who tested positive compared with 78% in 1995-6, 93% in 1996, and 95% in 1997), O153 in one case, and a mixed positive response for O2-O103 in one case. The O157 serogroup has predominated (191/369, 51%) in serology results in the past six years (figure 3).

In 1998, stool was cultured from 62 (83%) of the 75 patients about whom this information was provided. E. coli O157:H7 was identified in four of the 43 patients tested for it. Three of them had associated serology positive for serogroup O157, and one was not tested for O157. Fifty-seven per cent of specimens in 1998 were probed for E. coli O157 on MacConkey Sorbitol agar, a method recommended in most industrialised countries as the screening test for E. coli O157 infections, compared with 20% in 1997, but the diagnostic value of this method remains slight when done some time after the episode of prodromal diarrhoea: E. coli O157:H7 was identified in faeces in only three of 14 cases probed by stool culture and positive test results for serogroup O157. Other bacteria were identified in stool cultures from six other cases: campylobacter (2 cases), salmonella (1 case), yersinia in association with E. coli O128B12 (1 case), enterovirus echo 3 associated with an E. coli of uninterpretable serotype (1 case), and E. coli O125 (1 case)

In 1998, information about associated cases of diarrhoea was available for 69 out of 79 cases of HUS. Questioning of the entourage of HUS patients revealed cases of diarrhoea in two thirds (44/69) of the cases, 19 of whom were family members. Three episodes of cases of typical HUS clustered in space and time gave rise to further investigations.

In northern France, two distinct episodes of clustered cases led the regional health care authority (Direction Départementale des Affaires Sanitaires et Sociales, DDASS) and InVS to make enquiries. No dietary origin was found, but cases reported contact with patients with diarrhoea in the community in both episodes. The third episode was related to imported cases.

Six of the 85 cases reported by network paediatricians in 1998 occurred during or immediately after a stay abroad: two cases had returned from a west African country (Mauritania (1 case), Mali (1 case)) and four cases from a European Union country (Portugal (2 cases), Spain (1 case), Denmark (1 case)).

The two cases of HUS who had visited Portugal, reported the 27 October 1998 and 16 November 1998, were girls aged 3 and 2 years. They became ill two weeks apart; both had stayed during August in Villaverde (Portugal) but no other links (food, places of purchase, places of swimming) were found. Initial symptoms (prodromal diarrhoea) in the first case developed on 20 August 1998. HUS was diagnosed and treated two days later in Portugal. The second developed prodromal diarrhoea on 28 August 1998 and was diagnosed with HUS on 3 September 1998 at Rheims University Hospital. Serological testing, performed on one girl only, was positive for serogroup O157. The mother of one of the cases reported that several children in the community had attended the Villaverde dispensary because of gastroenteritis. InVS informed the Portuguese partner of the International Surveillance Network for Salmonella and VTEC O157 enteric infections (Enter-net) but received no information in response.

The proportion of imported cases (6/85) was much higher in 1998 than in previous years (1/94 in 1994, 2/94 in 1995, 1/100 in 1997). Stool specimens, taken from four cases, yielded only one culture of shigella, from the child returning from Mali, whose typing we have not received. Serological testing was done in four cases; only one of the two patients returning from Portugal was positive for serogroup O157. Enter-net reported no other case of HUS in 1998 that might have been related to a stay in these countries (6).

Cases of HUS occur sporadically in France. The incidence of paediatric HUS has been essentially stable for six years and remains similar to that found in other European countries (7,8).

Since the surveillance system was set up, 60% of cases of HUS tested serologically have been associated with a STEC infection, and serogroup O157 has been responsible for 92% of these infections.

The proportion of cases with positive results from serological tests has fallen steadily since 1995 (from 74% in the 1995-6 prospective study, to 66% in 1996, 51% in 1997, and 48% in 1998). In 1998 it was notably below the four year mean (62%). Recent publications confirm that most patients with typical HUS have high concentrations of LPS antibodies, whether or not STEC has been cultured (9). The steady fall in the proportion of positive test results (for 26 serogroups tested) since 1996 cannot be explained by a technical problem. Negative results could represent the absence of an immunological response of the patient to infection by an E. coli of a given serogroup, or a diagnostic test that does not contain the LPS of the serogroup responsible for the infection. These hypotheses could be tested if isolates of the strain of E. coli responsible for prodromal diarrhoea were to be available more often. It would then be possible to improve serological testing by including the LPS of the serogroups most often found to be responsible for typical HUS in France. Techniques for identification of STEC from the stools of patients need to improve, and serological tests that enable E. coli infection to be confirmed, regardless of the serogroup of the strain responsible, need to be developed.

Despite a notable increase in the frequency of probing for E. coli O157 on MacConkey Sorbitol agar when HUS is diagnosed, the sensitivity of this diagnostic method remains poor. This is probably because of the time elapsing between the onset of prodromal diarrhoea and the diagnosis of HUS (median 7 days), which prompts the test. Routine testing as part of the investigation of all cases of diarrhoea would probably enable more cases to be confirmed (10)

The HUS surveillance system set up with paediatric nephrologists is essential since it is, in France, the only tool currently available for monitoring the evolution of STEC infections and enabling the early detection of epidemics. The steady fall in the proportion of cases with positive serological test results suggests the hypothesis of evolution of new E. coli serogroups responsible for HUS and the possible emergence of new serogroups not included in diagnostic serology kits.

The high proportion of imported cases in France in 1998 emphasises the need to improve national and international early warning systems for these infections. Enter-net, as an international surveillance network, has an important part to play. One of its objectives is to detect clusters of STEC infections early in order to issue warnings swiftly to the countries concerned. It has been possible in this way to identify several international epidemics of E. coli O157/HUS (Denmark, England and Wales, Finland and Sweden) (11,12) that could not have been detected by national surveillance only.

These surveillance data were published previously in the Bulletin Épidémiologique Hebdomadaire 2000; 13: 55-8

1. Scottish Centre for Infection and Environmental Health. Late death arising from the central Scotland outbreak of infection with Escherichia coli O157:H7. SCIEH Weekly Report 1997; 31: 113.
2. Michino H, Araki K, Minami S, Takaya S, Sakai N, Miyazaki M, et al. Massive outbreak of Escherichia coli O157H7 infection in schoolchildren in Sakai City, Japan, associated with consumption of White Radish sprouts. Am J Epidemiol 1999; 150: 787-96.
3. Boyce TG, Swerdlow DL, Griffin PM. Escherichia coli O157H7 and the hemolytic-uremic syndrome. N Engl J Med 1995; 333: 364-8.
4. Decludt B, Bouvet P, Mariani-Kurkdjian P, Grimont F, Grimont PAD, Hubert B, et al. Haemolytic uraemic syndrome and Shiga toxin-producing Escherichia coli infection in children in France. Epidemiol Infect 2000; 124: 215-20.
5. Bastian SN, Carle I, Grimont F. Comparison of 14 PCR systems for the detection and subtyping of stx genes in Shiga-toxin-producing Escherichia coli. Res Microbiol 1998; 149: 457-72.
6. Fisher IST. The Enter-net international surveillance network – how it works. Eurosurveillance 1999; 4: 52-5.
7. Bitzan M, Ludwig K, Klemt M, Konig H, Buren J, Muller-Wiefel DE. The role of Escherichia coli O157 infections in the classical (enteropathic) haemolytic uraemic syndrome: results of a Central European, multicentre study. Epidemiol Infect 1993; 110: 183-96.
8. Ammon A. Surveillance of enterohaemorrhagic E. coli (EHEC) infections and haemolytic uraemic syndrome (HUS) in Europe. Eurosurveillance 1997; 2: 91-6.
9. Karch H, Bielaszewska M, Bitzan M, Schmidt H. Epidemiology and diagnosis of shiga toxin producing Escherichia coli infections. Diagn Microbiol Infect Dis 1999; 34: 229-43.
10. Tarr PI, Neill MA, Clausen CR, Watkins SL, Christie DL, Hickman RO. Escherichia coli O157:H7 and the hemolytic uremic syndrome: importance of early cultures in establishing the etiology. J Infect Dis 1990; 162: 553-6.
11. CDSC. European collaboration identifies an outbreak of Escherichia coli O157 infection in visitors to Fuerteventura, Canary Islands. Commun Dis Rep CDR Wkly 1997; 7: 127.
12. Pebody RG, Furtado C, Rojas A, McCarthy N, Nylen G, Ruutu P, et al. An international outbreak of Vero cytotoxin-producing Escherichia coli O157 infection amongst tourists; a challenge for the European infectious disease surveillance network. Epidemiol Infect 1999; 123: 217-23.