After a nursery school trip to a dairy farm, 20 (53%) of 38 children and 3 (23%) of 13 adult helpers developed gastrointestinal infection. Campylobacter jejuni was isolated from 15 primary cases and from 3 of 9 secondary household cases. A cohort study of the school party found illness to be associated with drinking raw milk (relative risk 5·4, 95 % confidence interval 1·4–20·4, P = 0·001). There was a significant dose response relationship between amount of raw milk consumed and risk of illness (X2-test for linear trend 12·1, P = 0·0005) but not with incubation period, severity of symptoms or duration of illness. All 18 human campylobacter isolates were C. jejuni resistotype 02 and either biotype I (number 16) or biotype II (number 2). Campylobacter was also isolated from samples of dairy cattle and bird faeces obtained at the farm but these were of different resisto/biotypes. Educational farm visits have become increasingly popular in recent years and this outbreak illustrates the hazard of exposure to raw milk in this setting.

A multiplex polymerase chain reaction (PCR) was developed to detect the presence of the ail, yst, and virF genes of Yersinia enterocolitica simultaneously, quickly and accurately. The amplified fragment sizes were 356 base-pairs (bp) for the ail gene, 134 bp for the yst gene, and 231 bp for the virF gene. The specificity of the amplified products was confirmed by hybridization with digoxigenin-labelled oligonucleotide probes. Amplification was successful whether the template was derived from a single colony of bacteria, aliquots of boiled bacterial suspensions, from DNA extracted from pure or mixed cultures or from stool specimens. Amplification of the virF gene was also achieved from strains of Y. pseudotuberculosis carrying the 70 kb plasmid but not with preparations from other related Yersinia species or from other members of the family Enterobacteriaceae. The detection limit we established was 5–10 colony forming units per millilitre (cfu/ml) and 1·0 pg of DNA.

Low-affinity penicillin-binding protein PBP-2a encoded by mecA is closely related to methicillin resistance in staphylococci, and expression of PBP-2a is controlled by regulator elements encoded by mecRl and mecI which are located adjacent to mecA on the chromosome. Deletion or mutation which occurred in mec regulator gene is considered to be associated with constitutive production of PBP-2a. The distribution of the mec regulator genes in 176 strains of Staphylococcus aureus and 33 strains of S. epidermidis isolated from a single hospital was studied by polymerase chain reaction amplification. Most clinical isolates of methicillin-resistant S. aureus (MRSA) (94.3 %) and S. epidermidis (MRSE) (83.9 %) possessed both mecI and mecR1 genes (type I), whereas no mec regulator genes were detected in mecA-negative isolates. In contrast, 7 MRSA and 5 MRSE isolates were found to have incomplete regulator genes, and they were classified into three groups; strains which lacked only mecI gene (type II), strains which lacked mecIand 3'-end of mecR1 gene (type III), and strains which lacked both regulator genes (type IV). Analysis of mecI gene from all the strains having mecI by restriction fragment length polymorphism after Mse I digestion indicated that three MRSA strains possessed one of the known point mutations identified previously. These findings indicated the predominance of a single type of MRSA possessing both mecI and mecR1 in the study period and also suggested a high genomic diversity in mec regulator region of staphylococci.

The significance of non-culturable forms of Campylobacter spp., especially with regard to the epidemiology of this organism in poultry flocks, was explored. Two different experiments were conducted to produce non-culturable Campylobacter spp. and test their ability to colonize the animal gut. In the first experiment a mixture of 28 different strains of Campylobacter spp. from various sources was inoculated in both sterilized surface water and potassium phosphate buffer and stored at 4 °C. After Campylobacter spp. were no longer detectable by culture in the microcosms, the mixtures of non-culturable cells were used to challenge both chicks and mice. Recovery of non-culturable Campylobacter spp. from the animals was not successful at 4 weeks after administration. In the second experiment the survival of six individual strains of Campylobacter spp. in sterilized surface water at 4 °C was studied and the resulting non-culturable cells were used to challenge chicks. None of the Campylobacter strains could be recovered from the chicks at 2 weeks after administration. We conclude that occurrence of non-culturable forms of Campylobacter spp. capable of colonizing chicks is not a common phenomenon and that non-culturable forms of Campylobacter spp. are likely to be insignificant for importantly to the epidemiology of the organism in Dutch broiler flocks.

During March–April 1993, an estimated 403000 residents of the 5-county greater Milwaukee, Wisconsin area developed cryptosporidiosis after drinking contaminated municipal water. Although the number of cases dropped precipitously after the implicated water plant closed on 9 April, cases continued to occur. To investigate risk factors for post-outbreak cryptosporidiosis, 33 Milwaukee-area residents who had laboratory-confirmed Cryptosporidium infection with onset of diarrhoea between 1 May and 27 June 1993 were interviewed by telephone. Of these, 28(85%) had onset of diarrhoea during May, 12(36%) had watery diarrhoea during the outbreak, and 5 (15%) were HIV-infected. In a neighbourhood-matched case-control study, immunosuppression (matched odds ratio (MOR) not calculable, 95 % confidence interval (CI) 3·0, infinity) and having a child less than 5 years old in the household (MOR = 17·0, CI 2·0, 395·0) were independently associated with infection. When persons who had diarrhoea during the outbreak were excluded, immunosuppression remained significantly associated with illness (MOR not calculable, CI 1·6, infinity). Cryptosporidium transmission continued after this massive waterborne outbreak but decreased rapidly within 2 months.

A collaborative exercise, supervised by the World Health Organisation, was set up to compare ELISAs used for the serological detection of Salmonella enteritica serotype Enteritidis in chickens. The aim was to ascertain how far agreement could be reached on the interpretation of optical density readings for high titre, intermediate titre and low titre sera. Two sets of sera were sent to 14 participants. The first set compared high, medium and low titre sera raised in specified-pathogen-free and commercial broiler breeder chickens. The second set comprised 20 sera of different antibody titres raised in commercial birds reared under laboratory conditions and sent blind. Both indirect and double-antibody sandwich blocking ELISAs were used with a number of different detecting antigens. With a few exceptions good agreement was reached on the interpretation of results obtained from high and low titre sera from the optical density obtained with a single serum dilution. Differences were observed in the interpretation of medium titre sera. The results suggested that most ELISAs produce reasonably comparable results and that practical problems may arise from interpretation of the results mainly as a result of the choice of the criteria used for differentiating sera obtained from infected and uninfected chickens. These problems are discussed.

A total of 279 Vibrio anguillarum strains were serotyped and examined for plasmid content. Plasmids were subjected to digestion with restriction enzymes. Most strains belonged to serogroup O1 (39%) and O2 (16%). In total 164 strains (53%) carried plasmids. Of the O1 and O2 isolates, 92% and 30%, respectively, carried one or more plasmids. Restriction fragment length polymorphism (RFLP) analysis of plasmid DNA indicated that plasmids belonged to several groups. Each group seemed to be restricted to a single O-serovar. The largest group was the pJM1-like plasmids among most serovar O1 strains. Most of these plasmids were about 67 kb like the pJM1 plasmid, but various derivatives ranged from 26–77 kb. RFLP studies of the 67 kb plasmids revealed 17 different restriction patterns. Some patterns were dominant among European strains whereas others were dominant among North American strains. The results confirmed the applicability of O-serotyping together with plasmid profile and restriction analysis of plasmids for typing of V. anguillarum. They also indicated that plasmids among strains which belonged to the traditional fish pathogenic serogroups, O1 and O2, showed more homology than did strains from most other serogroups, that were usually non-pathogenic, environmental bacteria.

A Register of HIV-infected persons who have had a negative antibody test within 3 years of their first antibody positive test (seroconverters) is being set up in the UK to monitor the distribution of times from HIV seroconversion to AIDS (the incubation period) and to death. It will also provide a national resource for use by those designing studies in this group of individuals. Clinicians caring for HIV-positive persons in Genito-Urinary Medicine, Infectious Disease and other departments throughout the UK were asked to participate by providing information on eligible subjects. Most laboratories undertaking HIV antibody testing were also contacted and asked to provide the name of the attending clinician for all seroconverters identified through the HIV laboratory reporting systems of the PHLS Communicable Disease Surveillance Centre (CDSC) and the Scottish Centre for Infection and Environmental Health (SCIEH) and for any other seroconverters known to them but not identified by CDSC or SCIEH. Data items sought for the Register include: sex, ethnic group, probable route of HIV transmission, annual CD4 counts, details of therapy and prophylaxis prescribed, AIDS-defining events and vital status. Follow up information is collected annually. Wherever possible, all seroconverters known to a clinic have been identified, whether currently alive or dead, either from clinic records or laboratory reporting or both. The objective is to establish and update a complete register of seroconverters on a long-term basis to provide reliable estimates of the incubation period on which future projections of AIDS cases in the UK can be made.

Two Enteritidis PT4 isolates which differed in inherent tolerance to heat, acid, H2O2 and the ability to survive on surfaces were used to infect mice, day-old chicks or laying hens. The acid-, heat-, H2O2- and surface-tolerant isolate was more virulent in mice and more invasive in laying hens, particularly in reproductive tissue. However, no significant differences were observed in behaviour in chicks. Both PT4 isolates were able to infect chicks housed in the same room as infected birds, although the heat-tolerant isolate survived significantly better than the heat-sensitive one in aerosols.

The incidence of acute gastroenteritis and self suspected food poisoning in general practice populations was compared with consultation rates in four group practices during a 3-month winter period and a 2-month autumn period. The average monthly consultation rate for acute gastroenteritis and self suspected food poisoning was 0·3 % and 0·06 % respectively. However, over the same period, on average, an estimated 7% of the practice population per month reported an acute gastroenteritis illness, and 0·7% suspected a food poisoning illness. Only about one in 26 people who suffer an acute episode of gastroenteritis consult their general practitioner (GP). In two practices, faecal samples were sought from all patient cases; the isolation rate for salmonellas was 2% (3/191) and for campylobacters it was 12% (23/191). In the other two practices following routine management, the isolation rate for salmonellas was 9% (6/64) and for campylobacters it was 2% (1/64). Isolation of faecal pathogens was not associated with patients' suspicion of food poisoning. A history of eating out in the week before onset was associated with a significantly increased yield of salmonellas and campylobacters.

A survey of 725 prepacked sandwiches was conducted examining for the presence on enrichment, and by plate count, of Listeria species. Sandwiches were found to contain Listeria sp. more frequently than their component foods. Chicken, beef and bacon fillings were associated with more frequent isolation. Salad also was associated with more frequent isolation, but the increase was not significant. On enrichment, over 15% of sandwiches contained Listeria species. L. innocua and L. monocytogenes were the only species isolated by plate count at numbers ≤ 100 cfu/g (1·5% of total samples). Potentially hazardous levels of L. monocytogenes (defined as ≤ 103 cfu/g) were found in two sandwiches examined (0·3%), indicating that although Total Viable Counts (TVCs) may often be high, the risk of listeriosis to vulnerable individuals from sandwiches is relatively low. It is importent to distinguish the risk of consuming potentially hazardous levels of a pathogen in food from the risk of contracting illiness as a result of such an event.

This paper uses meta-analysis of published data and a deterministic mathematical model of hepatitis B virus (HBV) transmission to describe the patterns of HBV infection in high endemicity areas. We describe the association between the prevalence of carriers and a simple measure of the rate of infection, the age at which half the population have been infected (A50), and assess the contribution of horizontal and perinatal transmission to this association. We found that the two main hyper-endemic areas of sub-Saharan Africa and east Asia have similar prevalences of carriers and values of A50, and that there is a negative nonlinear relationship between A50 and the prevalence of carriers in high endemicity areas (Spearman's Rank, P = 0·0086). We quantified the risk of perinatal transmission and the age-dependent rate of infection to allow a comparison between the main hyper-endemic areas. East Asia was found to have higher prevalences of HBeAg positive mothers and a greater risk of perinatal transmission from HBeAg positive mothers than sub-Saharan Africa, though the differences were not statistically significant. However, the two areas have similar magnitudes and age-dependent rates of horizontal transmission. Results of a simple compartmental model suggest that similar rates of horizontal transmission are sufficient to generate the similar patterns between A50 and the prevalences of carriers. Interrupting horizontal transmission by mass immunization is expected to have a significant, nonlinear impact on the rate of acquisition of new carriers.

In 1994 a Texas prison containing a population of mentally retarded inmates experienced a large tuberculosis outbreak. Fifteen cases of tuberculosis were identified (8 confirmed by positive cultures for Mycobacterium tuberculosis) and more than 100 inmates became infected. The culture-confirmed patients were infected with an identical strain of tuberculosis as demonstrated by polymerase chain reaction (PCR) based DNA fingerprinting technique. The prison followed standard tuberculosis infection control policies, but these controls were inadequate to prevent tuberculosis transmission in this special population. Two hundred and thirty inmates (119 inmates showing evidence of new tuberculosis infection or active disease and 111 healthy controls) were enrolled in the investigation. Inmate cell assignments, job duties, and educational classes were identified and medical chart reviews were conducted on all inmates. Tuberculosis transmission was associated with residing on the D Wing of the prison (OR = 25·84, P < 0·01), attending school in Classroom A (OR = 8·34, P = 0·01) and working on the prison utility work crew (OR = 2·52, P < 0·01). The index case in the outbreak had been prescribed 6 months of isoniazid (INH) chemoprophylaxis in 1988.

One hundred and fourteen cases of pneumococcal meningitis were identified by prospective laboratory based surveillance during 1990–3 in the North East Thames Region. Higher rates of disease were seen in Asians (2·1/100 000) than Caucasians (0·8/100 000) (P = 0·002). The incidence of meningitis was higher in children than adults, while mortality rates were highest in adults over the age of 60 (48%). In 72 cases, both blood and CSF were culture positive. Serotyping of 65 isolates collected identified 22 serotypes (and one non-typable) causing disease, the most common being serotype 6 (13 cases) and serotype 14 (11 cases). Overall, 90% of serotype antigens identified were represented in the 23 valent vaccine. Ribotyping of 62 isolates identified 35 different patterns, of which 26 were single types. Different ribotypes were found among isolates of the same serotypes, with the exception of serotype 14, where 9 of 11 isolates had the same ribotype pattern. Four percent of isolates had reduced susceptibility to penicillin, but no high level penicillin resistance was found.

There was a marked rise in scarlet fever mortality in England and Wales in the mid-nineteenth century and spectral analysis of the registration details, 1847–80, shows that the interepidemic interval was 5–6 years, but after 1880 the endemic level fell and the fatal epidemics disappeared. The dynamics of the scarlet fever epidemics can be represented by a linearized mathematical model and because the system is lightly damped, it could be driven by an oscillation in susceptibility. Epidemics were significantly correlated with dry conditions in spring/summer (P < 0·001), suggesting that these produced a low amplitude oscillation in susceptibility which drove the system. Epidemics also correlated (P < 0·001) with an oscillation in wheat prices but at a lag of 3 years, suggesting that malnutrition during pregnancy caused increased susceptibility in the subsequent children which interacted synergistically with seasonal dry conditions. Scarlet fever mortality was sharply reduced after 1880 in parallel with falling wheat prices suggesting that the remarkable period of high scarlet fever mortality (1840–80) was dependent on poor nutritive levels during that time.

To study the occurrence of hepatitis B virus (HBV) and hepatitis C virus (HCV) in patients with chronic liver disease (CLD) and hepatocellular carcinoma (HCC) in Pakistan, blood samples from 105 sequential patients with biopsy-proven CLD (n = 82) and HCC (n = 23) were tested for HBV and HCV markers. Of the 105, 87 (83%) had evidence of hepatitis B exposure, 58 (55%) were positive for hepatitis B surface antigen (HBsAg), 23 (22%) had hepatitis C antibodies and 25 (24%) had detectable HCV RNA. Significantly more patients with HCC had evidence of HBV exposure in the absence of HCV markers (49/82 vs. 20/23, odds ratio 4·49,95 % CI 1·17–25·16). The proportion of patients positive for HBsAg with no HCV markers was also significantly higher in the HCC group (34/82 vs. 18/23, odds ratio 5·08, 95% CI 1·59—18·96). There were more patients with only HCV markers in the CLD group than the HCC group but the difference was not statistically significant (19/82 vs. 1/23, odds ratio 6·63, 95% CI 0·93—288·01). A modified non-isotopic restriction fragment length polymorphism study on PCR products was used to investigate the epidemiology of HCV genotypes in Pakistan. Due to depletion of the initial samples, a second series of specimens collected one year afterwards was used. Fifteen out of 40 samples had amplifiable product and all were identified as type 3. A commercial serological typing method on the same samples also confirmed that type 3 was the predominant HCV genotype in Pakistan.

In a community hepatitis A outbreak in the Rochester, New York area, 64 of 79 (81 %) people with anti-hepatitis A 1gM-antibodies and onset of symptoms from 9 April–31 May 1994, recalled eating food obtained from a retail buyer's club. Eleven (65%) of 17 households with cases contained club members compared with 7 (21%) of 34 neighbourhood-matched control-households (matched odds ratio 8·5; 95% CI 1·7–41·6). Club employees who ate sugar-glazed baked goods were at fourfold increased risk for hepatitis. The source of infection was an 1gM- positive baker who contaminated baked goods while applying sugar glaze. Computer-generated purchase lists implicated 11–12 March and 21–24 March as the most likely dates when contamination occurred. This investigation demonstrates the importance of food workers adhering to established hygiene practices. Computer-generated commercial datasets can be useful in epidemiologic investigations.

During a prolonged outbreak of meningococcal disease caused by serogroup B serotype 15 sulphonamide-resistant strains in one British health district, there was considerable variation in attack rates by town. General practitioner (GP) antibiotic prescribing rates were compared in high and low incidence towns. The only significant difference found was that erythromycin prescribing was more frequent in the high incidence towns (rate ratio 4·0, 95% CI 3·2–4·8, in March 1987 and 3·0, 95% CI 2·4–3·7, in November 1987). This was probably due to increased GP consultation rates for upper respiratory tract infection (URTI), but higher erythromycin usage may have increased meningococcal acquisition rates or susceptibility to meningococcal disease. Antibiotic prescribing rate should be further investigated in defined areas of high and low incidence of meningococcal disease.

The macrophage infectivity potentiator gene (mip) from Legionella pneumophila is a major virulence factor of the species. Thus, mip-detection by amplification has been proposed to assess the presence of L. pneumophila in clinical and environmental samples. The distribution of mip-related sequences within the Legionellaceae was studied by DNA amplification using mip-specific primers followed by Southern blot hybridization with an internal probe. Thirty nine species (48 serogroups) of Legionellaceae were screened in this attempt. Using this approach, sequences related to mip were observed in 89% of the tested species including the most recently described L. fairfieldensis, L. lansingensis and L. shakespearei. In several cases, cloning and sequencing of the amplified products confirmed the high levels of similarity between the sequence found in non-pneumophila species with that of the L. pneumophila mip gene. This confirms previous reports that mip related genes are widespread among Legionellaceae and therefore specific detection of the species L. pneumophila cannot be based on mip-targeted amplification.

Although viral respiratory tract infections may predispose to meningococcal disease, strong evidence that they do so exists only for influenza. Data on laboratory reported cases of respiratory syncytial virus (RSV) infections and meningococcal disease in England and Wales from mid-1989 to mid-1994 were analysed. Although the rise in RSV cases preceded the rise in meningococcal disease cases each winter, the interval between the rise and fall of the two diseases was inconsistent, no association was found between time series after removal of the seasonal component, and there was no evidence that more cases of meningococcal disease occurred in winters with more RSV disease. RSV may have less effect on the two most likely mechanisms whereby influenza predisposes to meningococcal disease, namely lowered immunity and impaired pharyngeal defences.