Between November 1992 and February 1993, a large outbreak of Escherichia coli O157[ratio ]H7 infections occurred in the western USA and was associated with eating ground beef patties at restaurants of one fast-food chain. Restaurants that were epidemiologically linked with cases served patties produced on two consecutive dates; cultures of recalled ground beef patties produced on those dates yielded E. coli O157[ratio ]H7 strains indistinguishable from those isolated from patients, confirming the vehicle of illness. Seventy-six ground beef patty samples were cultured quantitatively for E. coli O157[ratio ]H7. The median most probable number of organisms was 1·5 per gram (range, <0·3–15) or 67·5 organisms per patty (range, <13·5–675). Correlation of the presence of E. coli O157[ratio ]H7 with other bacterial indicators yielded a significant association between coliform count and the presence of E. coli O157[ratio ]H7 (P=0·04). A meat traceback to investigate possible sources of contamination revealed cattle were probably initially colonized with E. coli O157[ratio ]H7, and that their slaughter caused surface contamination of meat, which once combined with meat from other sources, resulted in a large number of contaminated ground beef patties. Microbiological testing of meat from lots consumed by persons who became ill was suggestive of an infectious dose for E. coli O157[ratio ]H7 of fewer than 700 organisms. These findings present a strong argument for enforcing zero tolerance for this organism in processed food and for markedly decreasing contamination of raw ground beef. Process controls that incorporate microbiological testing of meat may assist these efforts.

The incidence of invasive meningococcal disease (IMD) among UK university students and non-students of similar age was investigated. In addition, we sought to identify structural risk factors associated with high rates of IMD in individual universities. Cases were ascertained via Consultants in Communicable Disease Control (or equivalent officers) between September 1994 and March 1997. Data on individual universities were obtained from university accommodation officers.

University students had an increased annual rate of invasive meningococcal disease (13·2/105, 95% CI 11·2–15·2) compared with non-students of similar age in the same health districts (5·5/105, CI 4·7–6·4) and in those health districts without universities (3·7/105, CI 2·9–4·4). This trend was highly significant. Regression analysis demonstrated catered hall accommodation to be the main structural risk factor. Higher rates of disease were observed at universities providing catered hall places for >10% of their student population (15·3/105, CI 11·8–18·8) compared with those providing places for <10% of students (5·9/105, CI 4·1–7·7). The majority of IMD amongst students was caused by serogroup B organisms.

University students in the UK are at increased risk of IMD compared with non-students of a similar age. The incidence of IMD tends to be greatest at universities with a high provision of catered hall accommodation.

In 1997 the first outbreak of Escherichia coli O157[ratio ]H7 infections involving 14 cases occurred in Finland. A case was defined as a resident of Alavus with an episode of diarrhoea between 5 and 17 July 1997, and from whom E. coli O157[ratio ]H7 was isolated from stool. The investigation included case searching and a population-based case control study. Five primary and eight symptomatic secondary cases of E. coli O157[ratio ]H7 illness were detected. In the 10 days before the outbreak, all 5 primary patients (aged 3–8 years), but only 6 of 32 population controls from the same age range (Fisher's test, P<0·001) and 4 of 10 sibling controls (P<0·05) had visited (but had not necessarily bathed in) a shallow beach popular among young children. Four out of 5 primary cases had remained within 5 m of the beach while swimming and had swallowed lake water compared to 1 of 5 population controls. These analytical epidemiologic findings incriminated fresh lake water as the vehicle of E. coli O157[ratio ]H7 transmission.

Potential risk factors for sporadic verocytotoxin-producing Escherichia coli (VTEC) infection in Belgium were investigated in a matched case-control study. Thirty-seven cases, 8 infected with O157 VTEC strains (all eaeA-positive), 29 with non-O157 VTEC strains (13 eaeA-positive and 16 eaeA-negative) and 69 matched controls were interviewed. In a conditional logistic regression analysis, consumption of fish appeared to be a risk factor for infection (adjusted odds ratio (OR) 3·25, P=0·04). Contact with dogs (OR 0·27, P=0·04) and consumption of shellfish (OR 0·19, P=0·05) showed a negative association, corresponding to a decrease in risk. These findings might be explained if low level environmental exposure to VTEC induces protective immunity. Eating raw meat, a frequent habit in Belgium, or hamburgers, or eating in a fast-food restaurant was not more frequently reported by cases than controls. The exposures causing sporadic infections with VTEC, in particular non-O157 strains, may be very different from those which led to outbreaks, and may account for more cases overall.

The purpose of this study was to determine the presence of thermophilic Campylobacter spp. and Salmonella spp. in sand from non-EEC standard and EEC standard designated beaches in different locations in the UK and to assess if potentially pathogenic strains were present. Campylobacter spp. were detected in 82/182 (45%) of sand samples and Salmonella spp. in 10/182 (6%). Campylobacter spp. were isolated from 46/92 (50%) of samples from non-EEC standard beaches and 36/90 (40%) from EEC standard beaches. The prevalence of Campylobacter spp. was greater in wet sand from both types of beaches but, surprisingly, more than 30% of samples from dry sand also contained these organisms. The major pathogenic species C. jejuni and C. coli were more prevalent in sand from non-EEC standard beaches. In contrast, C. lari and urease positive thermophilic campylobacters, which are associated with seagulls and other migratory birds, were more prevalent in sand from EEC standard beaches. Campylobacter isolates were further characterized by biotyping and serotyping, which confirmed that strains known to be of types associated with human infections were frequently found in sand on bathing beaches.

The distribution of 3001 cases of verocytotoxigenic Escherichia coli (VTEC) reported in the Province of Ontario, Canada, were examined to describe the magnitude of this condition geographically and to evaluate the spatial relationship between livestock density and human VTEC incidence using a Geographical Information System. Incidence of VTEC cases had a marked seasonal pattern with peaks in July. Areas with a relatively high incidence of VTEC cases were situated predominantly in areas of mixed agriculture. Spatial models indicated that cattle density had a positive and significant association with VTEC incidence of reported cases (P=0·000). An elevated risk of VTEC infection in a rural population could be associated with living in areas with high cattle density. Results of this study suggested that the importance of contact with cattle and the consumption of contaminated well water or locally produced food products may have been previously underestimated as risk factors for this condition.

We assessed the rate of salmonella infections and risk factors associated with infection in North East Thames in 1993. Cases of culture confirmed infection were identified through microbiology laboratories and environmental health officers in the North East Thames. A total of 1730 cases were reported and 209 of these individuals (those who could be contacted within a 3-week interval after onset of symptoms) and matched controls were interviewed by telephone. In addition randomly selected controls were interviewed over a 4-month period about recent gastric acid lowering medication and antimicrobial ingestion. Sixty-six serotypes were identified: S. enteritidis was isolated from 1179 (69%) cases, S. typhimurium from 221 (13%), S. virchow from 77 (4%) and S. newport 25 (1%). Infections were more frequent in summer months. Highest rates were documented in children under 2 years of age for S. enteritidis (108/100000) and under 1 year for S. typhimurium (36/100000). Using the Townsend score, highest isolation rates of S. enteritidis were in more prosperous areas (36/100000 vs. 27/100000; odds ratio (OR) 1·3, 95% confidence intervals (CIs) 1·2–1·6, P<0·0001), while for S. typhimurium, there was no relation between deprivation index and isolation rates areas (6·4/100000 vs. 6·1/100000; OR 1·1, 95% CIs 0·8–1·5, P=0·77). The case control study showed a significant association between ingestion of products containing raw eggs and S. enteritidis infection (8/111 cases vs. 0/110 controls; OR undefined, lower 95% CIs 3·4). Individuals with salmonella infection were significantly more likely to have travelled abroad in the week before the onset of illness [42/186 (23%) vs. 1/182 (0·5%) ; OR 40, 95% CIs=5·5–291, P<0·001] and to report gastroduodenal disease [11/143 (7 %) vs. 3/143 (2 %); OR 5·0, 95% CIs=1·1–23, P=0·04]. There was an association between illness and gastric acid-lowering medications [unmatched controls OR 22·3 (95% CIs 1·5–3·7, P=0·0002), matched controls OR 3·7 (95% CIs 1·0–3·8, P=0·07)], but no association with antimicrobial ingestion.

To estimate the incidence of Guillain-Barré syndrome (GBS) following Campylobacter jejuni infection (CI) we studied three populations where outbreaks of CI had occurred involving an estimated 8000 cases. No case of GBS was detected in the 6 months following the outbreaks in the local populations. The point estimate for the risk of GBS following CI estimated in this study was 0 in 8000 (95% confidence interval 0–3).

Three molecular typing methods, pulsed-field gel electrophoresis (PFGE), ribotyping, and flagellin (flaA) gene typing, were used to discriminate within a group of 28 Campylobacter jejuni, heat-stable serotype 55 (HS55) isolates derived from cases of campylobacter enteritis occurring throughout Scotland, including 9 isolates associated with an outbreak. PFGE was found to be most discriminatory, identifying 6 distinct profiles, followed by ribotyping (5 profiles), and then flagellin gene typing (4 profiles). The coincidence of all three genotypic markers identified a dominant clonal line within the HS55 group, accounting for each of the outbreak strains, and for 9 of the 19 sporadic strains. A second, closely related, clonal line accounted for a further 5 of the sporadic strains, and also included the HS55 reference strain. Identification and monitoring of such clonal lines should facilitate more effective future epidemiological surveillance of C. jejuni.

The plasmid pOG670, a 54 kb, conjugative plasmid that specifies resistance to ampicillin and kanamycin and belonging to the incompatibility group X (IncX), was transferred into 10 isolates of Salmonella enterica serotype Enteritidis belonging to 10 different phage types (PT1, 2, 3, 4, 8, 9, 9b, 10, 11 and 13). Acquisition of the plasmid by these strains did not result in the loss of any resident plasmids but resulted in phage type conversion in 8 of the 10 strains (PT1, 2, 4, 8, 9, 9b, 10 and 11). The observed changes in phage type were found to result from the loss of sensitivity to 3 of the 10 typing phages used (phages 3, 5 and 7). Where the conversion resulted in a change to a defined phage type, both the new and original PTs belonged to the same, previously described, evolutionary lines. Enteritidis PTs 1, 4 and 8, commonly associated with poultry world-wide, were converted to PTs 21, 6 and 13a respectively. The results indicate a different route for phage type conversion Enteritidis from others reported in the literature and, although IncX plasmids are not normally present in PT8 or PT13a, may suggest a possible mechanism/link connecting these phage types.

During July 1995, an outbreak of Salmonella typhimurium definitive type (DT) 170, an unusual strain, occurred in South Wales. A case-control study found that illness was associated with eating kebabs (odds ratio undefined, P=0·002), doner kebabs (odds ratio 7·9, 95% confidence interval 1·5–20·5, P=0·02) and kebabs with yoghurt based relish (odds ratio undefined, P=0·009) but not with eating kebabs with mayonnaise-based relish (odds ratio 2·4, 95% confidence interval 0·4–13·9, P=0·53). Environmental investigations discovered a complex web of producers and wholesale suppliers. Kebab meat and yoghurt had been supplied to the two main implicated outlets by a single wholesaler. Samples of raw minced lamb and several environmental swabs taken at the wholesaler were positive for S. typhimurium DT170. Blood- stained, unsealed yoghurt pots were observed to be stored under a rack of raw lamb. Investigators of food poisoning outbreaks linked to takeaway food should consider cross- contaminated relishes and dressings as well as undercooked meat as potential vehicles of infection.

In May 1996, the Georgia Division of Public Health was notified about a cluster of persons with Salmonella Enteritidis (SE) infections in Waycross, Georgia. A matched pair case-control study to determine risk factors for illness found a statistically significant association of SE infection with a history of having eaten at Restaurant A during the 5 days before onset of illness (relative risk = 13 [95% confidence interval (CI)=3–62, P<0·01]). In a second case-control study, to determine specific food exposures, consumption of a deep-fried Mexican dish (chile relleno) (4 of 21 cases vs. 0 of 26 controls, odds ratio undefined, 95% CI>1·46, P=0·034) was found to be significantly associated with SE infection. An environmental investigation found evidence of suboptimal food storage and cooking temperatures at Restaurant A; cross contamination of foods may have contributed to the low attributable risk identified for chile rellenos. Five of 37 Restaurant A food and environment specimens yielded SE strains. All five positive specimens were from chiles rellenos. Of the seven outbreak-associated strains (six patient isolates and one food isolate from Restaurant A) for which phage typing was conducted, all were phage type 34. A FDA traceback investigation through Restaurant A's single-egg supplier identified the potential source as three interrelated farms in South Carolina. Environmental culture from one of these farms yielded SE phage type 34. As a result of this outbreak, FDA helped institute a statewide egg quality-assurance programme in South Carolina to minimize SE contamination of eggs.

Laboratory-based surveillance of salmonella isolates serotyped at four state health departments (Illinois, Michigan, Minnesota and Wisconsin) led to the identification of multistate outbreaks of salmonella infections during 1990 (176 cases of S. javiana) and 1993 (100 cases of S. montevideo). Community-based case-control studies and product traceback implicated consumption of tomatoes from a single South Carolina tomato packer (Packer A) MOR 16·0; 95% CI 2·1, 120·6; P<0·0001 in 1990 and again in 1993 (MOR 5·7; 95% CI 1·5, 21·9; P=0·01) as the likely vehicle. Contamination likely occurred at the packing shed, where field grown tomatoes were dumped into a common water bath. These outbreaks represent part of a growing trend of large geographically dispersed outbreaks caused by sporadic or low-level contamination of widely distributed food items. Controlling contamination of agricultural commodities that are also ready-to-eat foods, particularly fruits and vegetables, presents a major challenge to industry, regulators and public health officials.

The influence of consumption of a lactic acid-fermented cereal gruel togwa with pH[les ]4 on the presence of faecal enteric bacteria such as campylobacter, enterohaemorrhagic Escherichia coli (EHEC[ratio ]O157), enterotoxigenic Escherichia coli (ETEC), salmonella and shigella was evaluated. Under 5 years old healthy children listed in an ascending order of age were alternatively assigned and given either a lactic-acid fermented cereal gruel togwa (test diet) or an unfermented cereal gruel uji (control diet) once a day for 13 consecutive days. The presence of the enteropathogens was examined in rectal swabs collected from the children at baseline (before feeding session started), on days 7 and 13, and additionally 14 days (follow-up day) after the feeding session had stopped. The swabs were cultured on to different optimal media for respective enteropathogen and confirmed by standard microbiological and serological methods. Campylobacter spp. dominated among the enteropathogens (62% out of total) followed by Salmonella spp., ETEC and Shigella spp. Children with isolated enteropathogens in the togwa group was significantly reduced (P<0·001) from 27·6% at baseline to 7·8, 8·2 and 12·7% on days 7, 13 and follow-up day, respectively. The effect was more pronounced in those children taking togwa >6 times during the study period. In the control group, there was a slight decrease from 16·7% at baseline to 11·4% on day 7 and 8·1% on day 13. On the follow-up day, enteropathogens were found in 22·6% of the children, which was significantly higher than in those children taking togwa >6 times. We conclude, that regular consumption of togwa with pH[les ]4, once a day, three times a week may help to control intestinal colonization with potential diarrhoea-causing pathogens in young children.

Results of a surveillance on cholera conducted with patients seen at the Children Hospital in Bangkok, Thailand from August 1993 to July 1995 are presented. Annually, isolation rates for Vibrio cholerae varied between 1·7 and 4·4% of patients with diarrhoea. V. cholerae O1 serotype Ogawa accounted for between 31 and 47% of patients cultured positive for V. cholerae, whereas the O139 serotype dominated in early 1994 after which it disappeared. Non-O1, non-O139 strains were isolated at similar rates as serotype O1 in 1993 and 1994, but accounted for 69% of V. cholerae culture positive specimens in 1995. However, the annual proportion of the isolation of non-O1, non-O139 strains showed little variation and remained low between 1·0 and 1·3%. Serotyping of 69 epidemiological unrelated non-O1, non-O139 strains produced 37 different O-serotypes. BglI ribotyping of serotypes containing more than two strains demonstrated a high degree of heterogeneity within and between serotypes, except seven serotype O37 strains which showed an identical ribotype suggesting clonality. None of the 69 strains hybridized with a cholera toxin probe and only two strains hybridized with a heat-stable enterotoxin probe. Susceptibility testing to 12 antibiotics showed that 40 of 69 (58%) non-O1, non-O139 strains were resistant to colistin, streptomycin and sulphisoxazole and 28 of 69 (41%) were multiple antibiotic resistant (MAR; [ges ]4 antibiotics). Although 26 of 69 (38%) strains contained one or more plasmids, the plasmids were of low molecular weights and did not seem to encode antibiotic resistance. The results of the present study showed that a high proportion of heterogenous MAR V. cholerae non-O1, non-O139 strains were isolated from children at the hospital. With reference to the emergence of V. cholerae O139 in 1992, we suggest that non-O1, non-O139 strains should be monitored carefully to detect new serotypes with a possible epidemic potential, but also to determine the development and mechanism of antibiotic resistance.

Certain fimbriae and the flagellae of Salmonella enterica serovar Typhimurium have been shown to contribute to attachment and invasion of gut epithelium in the murine typhoid infection model and to contribute to pathogenesis in the chick. However, little is known of the role these organelles play in Enteritidis poultry infections and, to study this, day-old chicks were dosed orally in separate experiments with defined multiply afimbriate and/or aflagellate mutant strains of Enteritidis. The colonization and invasion characteristics of each mutant were compared with those of the isogenic wild type strain by the determination of the number of bacteria recovered from livers and spleens at known time points post infection. Compared with wild type Enteritidis, a mutant unable to express flagella but retaining the genetic potential to express fimbriae was recovered post mortem from livers and spleens in significantly reduced numbers compared to the isogenic wild-type at all time points post infection (P<0·001). Conversely, a flagellate but multiply afimbriate mutant (defective for the elaboration of five different fimbrial types) and a flagellate but non-motile ‘paralysed’ mutant were recovered from livers and spleens in similar numbers to the wild-type. The data suggested that Enteritidis flagella, but not fimbriae, played an important role in pathogenesis in the chick model and that the flagellar apparatus itself and not motility per se contributed significantly to this role.

An internationally agreed and validated set of phages is used worldwide for the typing of strains of Staphylococcus aureus of human origin. However, because of the sometimes reduced susceptibility of methicillin-resistant strains (MRSA) to these phages, some of the national typing centres use locally isolated and characterized sets of experimental phages. In this trial, 42 such phages were distributed to 6 centres and tested against 744 isolates of MRSA with the intention of defining a phage set to augment the international set. The use of these experimental phages increased the percentage typability from 75% with the international set to 93% and the number of identifiable lytic patterns from 192 to 424. A subset of 10 experimental phages was selected. When this subset was compared with the experimental panel, the typability rate was 91% and 370 distinct patterns were obtained. This subset of phages has been distributed for international trial.

One-hundred-and-thirteen isolates of Salmonella serotype Thompson from diverse sources in seven countries were characterized by PvuII ribotyping and IS200 fingerprinting. Ten PvuII ribotypes were observed. The predominant PvuII ribotype 1 represented a major clone of world-wide distribution but was not found in Australia; PvuII ribotypes 2 and 3 represented minor clones. HincII ribotyping discriminated subtypes within PvuII ribotype 1: HincII ribotype 1 was distributed widely but HincII ribotype 2 was found mainly in Scottish isolates. None of 101 isolates of PvuII ribotypes 1–3 contained copies of IS200. All 12 isolates of PvuII ribotypes 4–10 were from Australia and 7 of them contained copies of IS200 of 5 different profiles. These results suggest the existence of at least two lineages of Salmonella Thompson with a different geographical distribution. The finding that most isolates from man and poultry in Scotland belonged to the same ribotype (PvuII 1/HincII 2) and were IS200-negative suggests that poultry is an important source of human infection in Scotland.

A pilot study to determine the microbiological quality of unpasteurized milk from goats and ewes sampled from farm shops, health food shops, and other retail premises found that 47%, (47/100) of goats' and 50% (13/26) of ewes' milk samples failed the standards prescribed by the Dairy Products (Hygiene) Regulations 1995. In addition, Staphylococcus aureus, haemolytic streptococci or enterococci, were present in excess of 102 c.f.u./ml in 9 (7%) 2 (2%) and 19 (15%) samples, respectively. Salmonella, campylobacter, verocytotoxin-producing Escherichia coli O157[ratio ]H7 and Listeria monocytogenes were not detected in the samples. At the time of purchase approximately half of the milk samples (58%) were frozen, the rest were liquid. Farm outlets sold predominantly liquid milk, other retail premises sold a frozen product. The microbiological quality of goats' and ewes' milk, whether frozen or liquid, was not significantly different. Milk sold from farm shops was of lower quality than that from health food shops and other retail premises. In this pilot study most producers (92%) supplied, and most retailers (76%) sold unpasteurized goats' and ewes' milk that contained unacceptable levels of indicator organisms. The study was carried out during the winter when goats' milk production is reduced. The results indicate the need for a full representative study of unpasteurized goats' and ewes' milk on retail sale throughout the year.

A prospective cohort study with 1 year follow-up evaluated the relation between MRSA carriage and mortality, likelihood of hospitalization and functional status in residents of a nursing home for the elderly. Included were all 447 residents living in the home in early June 1994. From all patients, swabs were taken from nose, throat and perineum. Additional swabs (sputum, urine or wounds) were taken when indicated. The relative risk (RR) of dying within 6 months in MRSA carriers compared to non-carriers was 2·29 (95% CI=1·04–5·04). This RR remained stable (1·57–2·40) after adjustment for co-variables using Mantel–Haenszel stratified analysis. After 1 year, the RR was reduced to 1·30 (95% CI=0·65–2·58). Univariate survival analysis confirmed a difference in survival between carriers and non-carriers after 6 months (log-rank P=0·04) and no difference after 1 year. Cox regression analysis resulted in a hazard ratio for dying within 6 months of 1·73 (95% CI=0·72–4·17). No relation was found between carriage and either likelihood of hospitalization or indicators of functional status. These results are compatible with a possible relation between 6 months mortality and MRSA carriage in nursing home patients. It calls for a large scale, multicentre cohort study in order to either confirm or refute these findings.