A large outbreak of Escherichia coli O157 infections via school lunches occurred at primary schools in 1996 in Sakai City, Japan. As many as 10000 patients suffered from diarrhoea, haemorrhagic colitis and haemolytic uraemic syndrome (HUS). Using data on 288 inpatient school children affected by this outbreak, of whom 36 presented complete HUS and the remaining 252 tested positive for E. coli O157 culture, we attempted to identify predictors for the progression to HUS. Within the first 5 days of illness, clinical features associated with inpatients who developed HUS compared with those without HUS included a C reactive protein (CRP) level higher than 1·2 mg/dl (OR 44·26; 95% CI 5·83–336·23), a white blood cell (WBC) count greater than 11·0 × 109/1 (OR 5·03; 95% CI 27·13–11·87) and a temperature higher than 38·0 °C (OR 5·00; 95% CI 2·25–117·08). It can be concluded that these three factors are predictive factors for the development of HUS in patients with E. coli O157 infection, and patients who have two or all of these factors should be observed closely.

In 1997, a community-wide outbreak of Salmonella enterica serotype Typhimurium (S. typhimurium) infection occurred in France. The investigation included case searching and a case-control study. A case was defined as a resident of the Jura district with fever or diarrhoea between 12 May and 8 July 1997, from whom S. typhimurium was isolated in stool or blood. One hundred and thirteen cases were identified. Thirty-three (83%) of 40 cases but only 23 (55%) of 42 community controls, matched for age and area of residence, reported eating Morbier cheese (Odds ratio: 6·5; 95% Confidence Interval: 1·4–28·8). Morbier cheese samples taken from the refrigerators of two case-patients and one symptom-free neighbour cultured positive for S. typhimurium of the same phage type as the human isolates. The analysis of distribution channels incriminated one batch from a single processing plant. These findings show that an unpasteurized soft cheese is an effective vehicle of S. typhimurium transmission.

Between December 1997 and January 1998 an increase in the number of isolates of Salmonella enterica serovar Newport, a serotype rarely causing indigenous infections in Finland, was detected. This included two clusters of gastroenteritis following funeral meals. An inquiry via Enter-net revealed a concomitant increase in cases of S. Newport in the United Kingdom. To investigate the Finnish outbreak, a total of 56 S. Newport strains (22 from the outbreak period, 27 from pre- and post-outbreak period, and 7 from imported food producing animals) were studied by pulsed-field gel electrophoresis (PFGE); selected isolates were also phage typed. Two retrospective questionnaire studies evaluating food exposures among the funeral attendants were conducted. All isolates from the clusters had an identical PFGE pattern which was also found in 13 infections temporally close to but not associated with the clusters. The Finnish outbreak was caused by the same phage type as the one in the United Kingdom. In both clusters, an epidemiological link between illness and exposure to cured ham was found. In conclusion, the outbreak was not limited to the two clusters but was more widely spread both in and outside Finland. Early alarm systems of food-borne outbreaks and collaboration between European countries are needed for investigating international outbreaks.

To understand the role of flagella and fimbriae of Escherichia coli O78[ratio ]K80 in avian colibacillosis, day-old chicks were dosed orally with defined afimbriate and or aflagellate mutants and colonization, invasion and persistence compared with that of the wild-type. In an invasion model, chicks were dosed with 1 × 105 c.f.u. of a single strain and mutants defective for type 1 fimbriae, curli fimbriae or flagella colonized livers by 24 h although the numbers of bacteria present were significantly less than the wild-type. Mutants colonized between 50 and 75% of spleens whereas the wild-type colonized 100% of spleens. Additionally, the numbers of mutant bacteria in colonized spleens were significantly less than the wild-type. Surprisingly, mutants defective for the elaboration of more than one appendage were no more attenuated than single mutants. In a persistence model, chicks were dosed with 1 × 102 c.f.u. of a single strain and mutants defective for type 1 or curli or flagella or any combination thereof persisted as assessed by cloacal swabbing for 5 weeks of the experiment less well than the wild-type. In an additional persistence model, chicks were dosed with 5 × 102 c.f.u. of each of wild-type and one mutant together. All mutants were significantly less persistent than the wild-type (P < 0·001) and one mutant which lacked type 1, curli and flagella, was eliminated within 2 weeks. Analysis of the trends of elimination indicated that flagella contributed to persistence more than curli, which contributed more than type 1 fimbriae. Here was evidence for a major role in colonization, invasion and persistence played by type 1, curli and flagella.

Thirty-nine strains of Salmonella typhi, isolated in 1995 from four Districts in Pakistan, Rawalpindi, Islamabad, Kharian and Jehlem, were catalogued and examined. Chromosomal DNA from each isolate was digested with XbaI restriction endonuclease and subjected to pulsed-field gel electrophoresis. Three clonal variants comprising of 17–19 DNA fragments were identified. Antibiotic susceptibility testing identified that 37 of the S. typhi were resistant to chloramphenicol, trimethoprim and ampicillin. These antibiotic resistance genes were found to be located on one of four plasmids belonging to incompatibility group IncHI1 and ranging in size from 150–175 Kb. The genes responsible for this resistance in each case were the chloramphenicol acetyltransferase (CAT) type I, the dihydrofolate reductase (DHFR) type VII and the β-lactamase TEM-1 respectively.

Salmonella Typhimurium definitive type 104 with chromosomally encoded resistance to five or more antimicrobial drugs (R-type ACSSuT+) has been reported increasingly frequently as the cause of human and animal salmonellosis since 1990. Among animal isolates from the northwestern United States (NWUS), R-type ACSSuT+ Typhimurium isolates increased through the early 1990s to comprise 73% of Typhimurium isolates by 1995, but subsequently decreased to comprise only 30% of isolates during 1998. NWUS S. Typhimurium R-type ACSSuT+ were consistently (99%) phage typed as DT104 or the closely related DTu302. S. Typhimurium isolates from cattle with primary salmonellosis, randomly selected from a national repository, from NWUS were more likely to exhibit R-type ACSSuT+ (19/24, 79%) compared to isolates from other quadrants (17/71, 24%; P < 0.01). Human patients infected with R-type ACSSuT+ resided in postal zip code polygons of above average cattle farm density (P < 0.05), while patients infected with other R-types showed no similar tendency. Furthermore, humans infected with R-type ACSSuT+ Typhimurium were more likely to report direct contact with livestock (P < 0.01) than humans infected with other R-types.

A model of the transmission of salmonella through the poultry meat production chain is developed, to predict the effects of intervention strategies for salmonella control. The model first describes the situation before intervention in terms of salmonella prevalences at flock level and some transmission parameters. After single control measures are translated into effects on these transmission parameters, the effects of sets of control measures (intervention strategies), can be calculated with the model. As research data are lacking, the model input parameters were derived from expert opinion. As an example, the effects of two intervention strategies proposed for the Dutch poultry industry are predicted. A sensitivity analysis is performed to indicate where the most effective control measures may be expected. Additionally, the reliability of the model predictions is studied by an uncertainty analysis. The use of the model as a tool for policy makers deciding about salmonella control strategies is discussed.

In this paper we report the distribution of Salmonella Typhi isolates in Italy and their resistance patterns to antibiotics. The data were collected by the Italian SALM–NET surveillance system in a pilot retrospective study of the period 1980–96. Data on drug-resistance were available for 82 isolates out of 176 S. Typhi isolated in Italy. Of these 82 isolates, 32 (39%) were resistant or intermediate to 1 or more antibiotics. Eight isolates were resistant and 7 intermediate to streptomycin; 4 isolates were resistant to ampicillin alone or in association with other antibiotics; only 2 strains (1 isolated in Lombardia in 1993 and the other 1 in Lazio in 1994) were resistant to chloramphenicol, and 2 (isolated in Sardegna and Piemonte in 1995 and 1996, respectively) showed intermediate resistance to chloramphenicol. The strains showing resistance to 3 or more antibiotics were very scarce: 1 (with 5 complete resistances) was isolated in Lazio in 1994, and another 1 (with complete resistance to 10 antibiotics and intermediate resistance to 2 antibiotics) was isolated in Molise in 1988. In conclusion, besides the routine activities to control typhoid fever, an accurate and continuous surveillance is necessary in order to quickly identify multidrug-resistant (MDR) S. Typhi strains and prevent their spread, even though their level, in our country, is still quite low.

Vellore is endemic for cholera due to Vibrio cholerae O1 and O139. In a previous study the prevalence of Vibrio cholerae in drinking water, lakes and sewage outfalls in a single 2-months period in Vellore, India was determined. In addition water samples from three sites were also tested for the presence of V. cholerae O1 and O139 by fluorescent antibody staining. This follow on study has examined how the environmental distribution of V. cholerae at the same sites alters over a 12-month period and the relationship to the clinical pattern of cholera in Vellore. Samples of water were collected from fixed sites at three water bodies each month between April 1997 and March 1998. Bacteria isolated from samples were identified by standard biochemical tests and isolated strains of V. cholerae tested for their ability to agglutinate O1 and O139 antisera. Samples were also tested for the presence of V. cholerae O1 and O139 by fluorescent antibody staining. The clinical isolation rate of V. cholerae in Vellore, maximum temperature and rainfall were also studied. The results demonstrate the presence in the environment of viable but non-cultivable (VNC) V. cholerae in 10 of 12 months of the study year as well as their viability. Their prevalence in the environment also correlated with the isolation of these pathogens from clinical samples over the same study period.

Between April 1996 and March 1997 we examined 5093 samples of raw beef and lamb products for the presence of E. coli O157. Samples were purchased from 81 small butchers' shops in south Yorkshire. In March 1997 we also examined five samples of dried mint for the presence of E. coli O157.

Strains of E. coli O157 were isolated by enrichment culture in modified buffered peptone water followed by immunomagnetic separation and culture of magnetic beads onto cefixime tellurite sorbitol MacConkey agar. Strains were characterized by phage typing, toxin genotyping and plasmid analysis.

Strains of E. coli O157 were isolated from 72 (1·4%) of 5093 samples; it was isolated from 36 (1·1%) of 3216 samples of beef products and from 29 (2·9%) samples of lamb products. The highest prevalence was found in lamb sausages and lamb burgers where E. coli O157 was isolated from 3 (4·1%) of 73 and 18 (3·7%) of 484 samples respectively. Strains of E. coli O157 were isolated most frequently during early summer. Strains of E. coli O157 were also isolated from 2 of 5 samples of dried mint although we did not determine how the mint had become contaminated.

All isolates of E. coli O157 were Verocytotoxin-producing as determined by both Vero cell assay and DNA hybridization for the genes encoding Verocytotoxin and all were positive for the eaeA gene. A combination of phage typing, toxin genotyping and plasmid profile subdivided the 72 strains of E. coli isolated into 20 different subtypes, of which 18 were indistinguishable from strains isolated previously from cattle and sheep; of these 18 strains, 8 were indistinguishable from strains isolated from human cases of infection during the study period.

Forty-three clinical strains of V. cholerae O1 biotype El Tor were isolated between 3 May and 10 June 1998 during an outbreak in the metropolitan area of Kuala Lumpur and its suburbs. With the exception of three Inaba strains that were restricted to three members of a family, all the others belonged to the Ogawa serotype. The strains were analysed for clonality using ribotyping and pulsed-field gel electrophoresis (PFGE). Two ribotypes, V/B21a and B27, were identified among 40 Ogawa isolates using BglI restriction endonuclease. Ribotype V/B21a has been described previously from Taiwan and Colombia and several Asian countries while B27 has been reported among isolates from Senegal. The three Inaba strains belonged to one ribotype, designated type A, not previously reported. PFGE analysis using NotI revealed that all isolates within a ribotype had identical profiles demonstrating clonality amongst the strains. Dice coefficient analysis of the two Ogawa genotypes revealed 89% similarity on ribotype patterns and 91.3% on PFGE profiles. Ribotype V/B21a isolates were associated with cases from dispersed areas of Kuala Lumpur and its suburbs while ribotype B27 was restricted to cases from one particular area suggesting a common-source outbreak.

One-hundred and twenty-one isolates of Salmonella enterica serotype Montevideo, representing different biotypes and incidents of infection detected in the UK between 1977 and 1995, were analysed by EcoRI ribotyping, PvuII ribotyping and IS200 fingerprinting. Among the isolates examined, 7 EcoRI ribotypes, 5 PvuII ribotypes and 55 IS200 profile types were recognized and 4 arbitrary groups defined. All 33 isolates of biotype 2d belonged to EcoRI/PvuII ribotype 1/1 and IS200 lineage A and comprised Group I. The other 88 isolates of biotype 10di and its variants were assigned to Groups II–IV. All 27 isolates in Group II were of EcoRI/PvuII ribotype 2/2 and IS200 lineage B. Among the 43 isolates in Group III, 42 of which were of EcoRI/PvuII ribotype 3/3, IS200 analysis identified 38 profiles in lineages C–I. Six EcoRI/PvuII ribotypes and 8 IS200 profiles, mostly in lineages C–E, were recognized among the 18 isolates in Group IV. The combined use of biotyping and ribotyping, and to some extent IS200 profiling, has enhanced our understanding of the clonal structure of serotype Montevideo and provides a basis for further study.

We conducted a study to determine the incidence of haemolytic uraemic syndrome (HUS) in children in France and to assess the role of Shiga-toxin-producing Escherichia coli (STEC) infection in the aetiology of HUS. In collaboration with the Société de Néphrologie Pédiatrique we undertook a retrospective review of all cases of HUS hospitalized from January 1993 to March 1995 and a 1-year prospective study (April 1995–March 1996) of epidemiological and microbiological features of cases of HUS. The polymerase chain reaction (PCR) procedure was used to detect stx, eae, e-hlyA genes directly from case stool samples. Serum samples from cases were examined for antibodies to lipopolysaccharide (LPS) of 26 major STEC serogroups. Two hundred and eighty-six cases were reported. The average incidence per year was 0·7/105 children < 15 years and 1·8/105 children < 5 years. During the prospective study, 122/130 cases were examined for evidence of STEC infection using PCR and/or serological assays and 105 (86%) had evidence of STEC infection. Serum antibodies to E. coli O157 LPS were detected in 79 (67%) cases tested. In conclusion, this study showed that STEC infection is an important cause of HUS in children in France, with a high proportion related to the O157 serogroup.

Human salmonellosis in Germany has been an increasing problem since the 1980s, with a peak of 195000 reported cases in 1992. During the peak years, isolates of Salmonella Enteritidis (SE) predominated by far over other salmonella serovars (NSE) (80 vs. 20%). In a comparison of the clinical characteristics of 790 persons infected with SE to 175 persons infected with NSE, watery diarrhoea (OR 1·7) and high grade (> 39 °C) fever (OR 1·8) were independently associated with SE infection. When comparing possible risk factors for acquiring salmonella infection among patients with SE compared to those with NSE, consumption of raw eggs (OR 4·4; P = 0·0006) was the most significant alimentary risk factor for SE infection, while travel outside Europe was negatively associated with SE infection (OR 0·08; P = 0·0001). When comparing all patients with salmonella infection, regardless of serovar, with healthy controls, consumption of raw eggs (OR 30·3; P = 0·001), of raw or undercooked eggs (OR 1·9; P = 0·003), or having puppies, kittens or turtles (OR 6·8; P = 0·002), were risk factors for salmonellosis.

During October 1996, an outbreak of Escherichia coli O157[ratio ]H7 infections among Connecticut residents occurred. An epidemiologic investigation included enhanced surveillance and a case-control study. Clinical isolates of Escherichia coli O157[ratio ]H7 were typed by pulsed-field gel electrophoresis (PFGE). Implicated cider samples were analysed by culture and polymerase chain reaction (PCR). Consumption of implicated cider was associated with illness; (matched odds ratio = undefined, 95% confidence interval = 3·5–infinity). Ultimately, a total of 14 outbreak-associated patients were identified. All isolates analysed by PFGE yielded the outbreak-associated subtype. Escherichia coli O157[ratio ]H7 was not cultured from three cider samples; PCR analysis detected DNA fragments consistent with Escherichia coli O157[ratio ]H7 in one. This outbreak was associated with drinking one brand of unpasteurized apple cider. PFGE subtyping supported the epidemiologic association. PCR analysis detected microbial contaminants in the absence of live organisms. Washing and brushing apples did not prevent cider contamination.

We examined the incidence rate and prognosis of tuberculosis in a cohort of patients with liver cirrhosis in Denmark. In a study cohort of 22675 patients with liver cirrhosis, we identified 151 cases of tuberculosis from 1977 to 1993. The incidence rate was 168·6 per 100000 person-years of risk, and the highest incidence rate was among men above 65 years of age, with an incidence rate of 246·0 per 100000 person-years of risk. The 30-day case-fatality rate was 27·3% and the 1-year case fatality rate was 47·7%. The results demonstrate that patients with liver cirrhosis are at increased risk of tuberculosis. Additionally, it is suggested that liver cirrhosis is an independent risk factor for tuberculosis, and that patients with liver cirrhosis who acquire tuberculosis have a poor prognosis.

Antimicrobial susceptibilities of Vibrio cholerae strains isolated from cholera patients admitted to the Infectious Diseases Hospital, Calcutta, India for 6 years were analysed to determine the changing trends; 840 V. cholerae strains isolated in 1992–1997 were included in this study. Among V. cholerae serogoup O1 and O139, ampicillin resistance increased from 1992 (35 and 70%, respectively) to 1997 (both serogroups 100%). Resistance to furazolidone and streptomycin was constantly high among V. cholerae O1 strains with gradual increase in resistance to other drugs such as ciprofloxacin, co-trimoxazole, neomycin and nalidixic acid. V. cholerae O139 strains exhibited susceptibilities to furazolidone and streptomycin comparable with those of O1 strains. However, after initial increase in resistance to chloramphenicol and co-trimoxazole, all the V. cholerae O139 strains became susceptible to these two drugs from 1995 onwards. Both V. cholerae O1 and O139 remained largely susceptible to gentamicin and tetracycline. V. cholerae non-O1, non-O139 strains, in contrast, exhibited high levels of resistance to virtually every class of antimicrobial agents tested in this study especially from 1995. Kruskal–Wallis one-way analysis showed that V. cholerae O1 Ogawa serogroup exhibited significant yearly increase in resistance to nine antibiotics followed by non-O1 non-O139 and O139 strains to six antibiotics and two antibiotics respectively. Interesting observation encountered in this study was the dissipation of some of the resistant patterns commonly found among V. cholerae non-O1 non-O139 or O1 serogroups to the O139 serogroup and vice versa during the succeeding years.

We identified a 5-year outbreak of a methicillin-susceptible Staphylococcus aureus (MSSA) strain, affecting 202 babies on a neonatal unit, by routine weekly phage typing all S. aureus isolates. Multiple staged control measures including strict emphasis on hand hygiene, environmental and staff surveillance sampling, and application of topical hexachlorophane powder failed to end the outbreak. S. aureus PT 53,85 (SA5385) was found on opened packs of Stomahesive®, used as a neonatal skin protectant. Only following the implementation of aseptic handling of Stomahesive®, and the use of topical mupirocin for staff nasal carriers of SA5385, and for babies colonized or infected with S. aureus, did the isolation rate of SA5385 decline. DNA fingerprinting indicated that [ges ]95% of SA5385 isolates were clonal. In vitro death rates of SA5385 on Stomahesive® with human serum were significantly lower than on Stomahesive® alone (P = 0·04), and on cotton sheet with serum (P = 0·04), highlighting the potential of this material as a survival niche. Phage typing remains a valuable, inexpensive and simple method for monitoring nosocomial MSSA infection.

We examined three populations from the Tehran region and the North part of Iran (Gilan), in all more than 5000 individuals, for carriage of Streptococcus pyogenes (group A streptococci; GAS). Children or adults with acute pharyngotonsillitis and healthy school children harboured GAS in 34·1, 20·0 and 21·0%, respectively. Typing of 421 randomly selected isolates showed a predominance of M-types M4, M5, M11, M12, as well as the provisional type 4245; however, many of the isolates were T and M non-typable. Forty-three percent of all strains were opacity factor (OF) negative. The type distribution differed markedly from that reported in 1973–4, when M types 1 and 12 were predominant.

The occurrence of extended-spectrum β-lactamase producing enterobacteria (ESBLE) has been prospectively surveyed in a neurosurgical intensive care unit (ICU). Of the 47 patients examined, 8 were identified as faecal carriers, and 2 of them developed a subsequent urinary tract infection. ESBLE were also detected in the immediate environment of five colonized and/or infected patients. All isolates were Klebsiella pneumoniae of a particular biotype which exhibited a similar antibiotype and produced an SHV-4 type β-lactamase. However, plasmid profiling and ribotyping revealed that strains isolated from seven patients of hall A were a single epidemic clone, whereas strains isolated from the eighth patient of hall B were different. Comparison between the characteristics of patients who carried an ESBLE during the surveillance period, and control patients who did not, showed that a recent surgery, and the length of ICU stay were significantly associated with the acquisition of ESBLE.