A model system for the study of phage conversion of Salmonella enterica serotype Enteritidis is reported. Temperate phages 1,2,3 and 6 from the phage typing scheme were used to convert several individually recognized phage types into others. Phage type 4 was converted to PT8, PT6a to PT4, PT6a to PT7, PT13 to PT13a and PT15 to PT11; some new phage lysis patterns were also detected.

This model was used to examine the relationships between phage types within a previously denned clonal lineage, SECLIII, to establish whether or not Enteritidis like Salmonella enterica serotype Typhi and Salmonella enterica serotype Paratyphi B possessed type determining phages. We were able to convert PT1 to PT20, and PT15 to PT11.

Molecular variation within and between plasmids of Salmonella enterica serotype Dublin was analysed. Such variation has been demonstrated in the serotype-specific plasmids (SSP's) of Typhimurium and Enteritidis. The two aims of this study were to determine the plasmid diversity in a host-adapted serotype and also the incidence of molecular variation in the SSP among strains of Dublin using restriction endonuclease fragmentation pattern (REFP) analysis with Pst1, Sma1 and EcoRV. Sixty-five strains were examined from seven countries. Plasmid profile and REFP analysis showed that none of the strains was plasmid-free. Seventy-seveń percent of the strains possessed the 72 kb SSP either alone or in combination with another plasmid; 23 % harboured plasmids which were molecular variants of the SSP. Four of the variants were more closely related to each other than to the reference SSP and were harboured by Dublin isolated from both the USA and Europe. A further three were shown to be cointegrate plasmids and were similarly distributed. Thirty-two percent of strains possessed the SSP alone. None of the UK strains was resistant to any of the antimicrobial agents tested whereas 74 % of the remaining strains were resistant to between one and five antimicrobial agents. This study corroborates previous findings concerning the high degree of stability of the SSP and confirmed the clonal nature of Dublin. Co-resident plasmids provided evidence of sub-clones within localized geographical areas.

In response to the Latin American cholera epidemic, El Salvador began a prevention programme in April 1991. The first case was confirmed in August, and 700 cases were reported within 3 months. A matched case-control study was conducted in rural La Libertad Department in November 1991. Illness was associated with eating cold cooked or raw seafood (odds ratio [OR] = 7·0; 95% confidence limits [CL] = 1·4, 35·0) and with drinking water outside the home (OR = 8·8; 95% CL = 1·7, 44·6). Assertion of knowledge about how to prevent cholera (OR = 0·2; 95% CL = 0·1, 0·8) and eating rice (OR = 0·2; 95% CL = 0·1, 0·8) were protective. More controls than patients regularly used soap (OR = 0·3; 95% CL = 0·1, 1·0). This study demonstrated three important points for cholera prevention: (1) seafood should be eaten cooked and hot; (2) populations at risk should be taught to treat household drinking water and to avoid drinking water outside the home unless it is known to be treated; and (3) education about hygiene can be an important tool in preventing cholera.

Extensive DNA sequence diversity was noted in Helicobacter pylori flagellin genes flaA and flaB. PCR amplified sequences from 49 isolates were digested with AluI. HindIII. MboI or MspI, the resultant patterns were compared between the different isolates and these used to differentiate the isolates from each other. Evidence that the extensive diversity that was found in these genes is the result of reassortment of sequences between strains in the bacterial population is presented, such that a comparatively small number of individual sequence mutations can recombine together in random combinations to form a greater number of distinct alleles. Geographical differences in the predominant patterns in the flaA alleles were also observed and could reflect regional differences either in the human host population or in the bacterial population. In view of the genetic complexity of this species, molecular typing schemes designed to identify related strains may falsely associate strains if the methods do not characterize sufficient genetic sites to exclude chance associations of genetic markers in strains which are actually not closely related to each other.

We studied 7 patients with nosocomial Legionnaires' disease to determine the relationship between isolates of Legionella pneumophila recovered from potable water and those recovered from patients. Potable water was cultured from all rooms in which patients had stayed prior to the diagnosis of Legionnaires' disease. The 38 isolates of L. pneumophila (31 environmental, 7 patient) were resolved into 9 distinct patterns by pulse-field gel electrophoresis (PFGE), 3 by plasmid content and 2 each with monoclonal antibodies and conventional agarose gel electrophoresis of small fragments of DNA.

Using PFGE it was determined that 4 of the 7 patients were infected with L. pneumophila identical to an isolate recovered from the potable water supply in one of the rooms each had occupied prior to the diagnosis of Legionnaires' disease. Patients had resided in a mean of 3·57 rooms before a diagnosis of nosocomial Legionnaires' disease. We conclude that in the setting of contaminated potable water and nosocomial Legionnaires' disease water from all the rooms which the patient has occupied prior to this diagnosis should be cultured. PFGE of large DNA fragments discriminated best among the isolates of L. pneumophila.

Pilot studies to test methods to determine the incidence, agents, risk factors and socioeconomic costs of infectious intestinal disease (IID) in England were carried out as recommended by the Committee on the Microbiological Safety of Food (the Richmond Committee) by eight general practices. There were case control and enumeration studies of patients presenting to general practice with IID, a population-based prospective cohort study, and a survey of socioeconomic costs of cases of IID. Information on risk factors was obtained by questionnaire (self-administered compared with interview) and a stool sample was requested on all cases and controls. Response rates in the GP case control study were 75% for case questionnaires and 74% for stools; for controls the figures were 70% and 68% respectively. The acceptance rate into the cohort study was 49%; this was significantly higher where phone contact was made. The rate was similar if recruitment was by individual or household. Follow-up of the cohort by negative reporting was complete for up to 6 months. Direct postage by subject was required to obtain fresh stool specimens. Estimates were obtained of presentation rates of IID and the distribution of risk factors which were used to plan the main study. The pilot study demonstrated that it is possible to undertake a national study based in general practice to determine the incidence of IID in the population and presenting to GPs and its agents, risk factors and costs.

Following an episode of water contamination with sewage in a rural Irish town, a community-wide survey of gastrointestinal-associated illness and health service utilization was conducted. Random sampling of households yielded residents who were surveyed using a self-administered questionnaire. Of 560 respondents from 167 (84%) households, equal proportions lived in areas known to have been exposed and unexposed to the contaminated water, although 65% of subjects reported using contaminated water. Sixty-one percent of subjects met the case definition. The most common symptoms among cases were abdominal cramps (80%), diarhoea (75%), appetite loss (69%), nausea (68%) and tiredness (66%). Mean duration of illness was 7·4 days. Only 22 % of cases attended their general practitioner. Drinking unboiled water from the exposed area was strongly associated with being a case. A substantial degree of community illness associated with exposure to contaminated water was observed. The episode ranks as one of the largest reported water-borne outbreaks causing gastrointestinal illness in recent times.

A prospective longitudinal survey of cutaneous leishmaniasis (Leishmania peruviana) was carried out in Peru on a study population of 4716 persons living in 38 villages (Departments of Lima, Ancash and Piura). Demographic and clinical data were collected from all individuals, and a Montenegro skin test (MST) was carried out on 72% (3418) of the study population. Each household was revisited at 3-monthly intervals for up to 2 years to detect new leishmaniasis cases; 497 people received a second MST at the end of the study. Analysis of the epidemiological data indicated that (i) 17% (16/94) of all infections were subclinical, (ii) this percentage increased significantly with age, (iii) clinical infections led to 73·9 % protective immunity (95 % C.I. 53·0–85·5 %) and relatively permanent MST responsiveness (recovery rate = 0·0098/year; 95% C.I. 0·000–0·020/year), (iv) sub-clinical infections led to protective immunity, which was positively correlated with their MST induration size (increasing by 17·9% per mm; P < 0·0001), and a mean MST recovery rate of 0·114/year (4/421 man-months), and (v) recurrent leishmaniasis was dominated by reactivations, not by reinfections.

Measles incidence in England and Wales has fallen to an all-time low. Attention is now focused on preventing local outbreaks, and, in the long run, on the elimination of indigenous measles. A realistic age-structured (RAS) mathematical model of measles transmission is used to reconstruct the impact of measles vaccination in England and Wales from 1968 to the present and to evaluate the merits of future policy options. In general, the predictions of the model show good agreement with long-term age stratified case reports and seroprevalence surveys. The model underestimates the proportion of cases that are notified in 0–2-year-old children. However, recent work suggests a high degree of misdiagnosis in this age group. Projections on the basis of the existing vaccination strategy in the UK suggest that the present level of measles vaccine coverage will be insufficient to eliminate small seasonal outbreaks of measles. This result is, however, sensitive to the assumed level of vaccine efficacy. Explorations of a variety of changes to current vaccination strategy favour a 2-dose schedule with the second dose administered at age 4 years irrespective of vaccination history. A vaccination campaign in school-age children, to reduce deficits in herd immunity, would accelerate progress towards measles elimination.

We analysed nation-wide reported measles cases during the 1988–9 epidemic and found that longer duration and wider spread were two major characteristics of the outbreak. All the 22 county/city index cases were reported with a delay of > 4 days and 64% were aged 5–14 years. This epidemic occurred mainly among 5–14-year-old school-children (59%), infants under 1 year (19%), and pre-school children (18%). The overall attack rate was 0·63 cases per 10000 population, with the highest attack rate (7·4 cases per 10000 population) occurring in infants. Among 280 confirmed cases <15 months of age, 9-month-old infants (42 cases) had a higher risk of measles and peaked at 10 months (49 cases). This epidemic started in March 1988 among 5–9-year-old children in the northern suburban area, then spread to Taipei City and neighbouring counties or cities. It continued to spread from the northern to southern and western areas during the summer vacation and New Year holidays. Multiple logistic regression analysis showed that the delay of measles reporting was strongly associated with the cases infected early in the epidemic (OR= 6·9, P < 0·001) and reported from teaching hospitals (OR = 2·6, P < 0·001). The reappearance of high attack rates among 5–9-year-old children in the 1985 and 1988–9 measles epidemics in Taiwan indicated the persistence of pockets of susceptible individuals even after mass immunization.

The effect of prior infection with cytomegalovirus (CMV) on progression of HIV disease in a cohort of 111 men with haemophilia was studied after 13 years followup. The relative hazards associated with CMV positivity on progression to AIDS, death and a CD4 count of 0·05 × 109/1 were 2·28, 2·42 and 2·34, respectively. CMV seropositive patients were significantly older than the seronegative and this was controlled for by using a Cox proportional hazards model. The relative hazards for the three endpoints decreased to 1·89, 1·82 and 1·93 respectively and were marginally non-significant (P = 0·05, 0·08 and 0·08 for the three endpoints respectively). We conclude that this cohort continues to show evidence of a ‘co-factor’ effect associated with prior infection with CMV which is confounded by age but not completely explained by age differences. The potential biological significance of these results is discussed in the context of recent controlled clinical trials which show a survival benefit from long-term high-dose acyclovir, a drug with activity in vivo against CMV and other herpesviruses.

A serological survey for the prevalence of hantavirus infections in The Netherlands was carried out on > 10000 sera, from selected human populations, and different feral and domestic animal species. Hantavirus-specific antibodies were found in about 1% of patients suspected of acute leptospirosis, 10% of patients with acute nephropathia, and in less than 0·1% haemodialysis and renal transplant patients. Among individuals with a suspected occupational risk, 6% of animal trappers, 4% of forestry workers, 2% of laboratory workers and 0·4% of farmers were seropositive. The majority of the seropositive individuals lived in rural and forested areas. The main animal reservoir of the infection was shown to be the red bank vole (Clethrionomys glareolus). Epidemiological, clinical and laboratory findings seen in serologically confirmed human cases were similar to those associated with nephropathia epidemica.

Astrovirus serotype 4 has increased in relative prevalence in the Oxford, UK area in 1993. The structural gene of human astrovirus serotype 4 has been sequenced and the results indicate that this protein differs substantially from serotypes 1 and 2. In particular, conservation at the C terminus is greatly reduced. However, amino acid substitutions in this region show a strong conservation in character suggesting that structural or functional constraints operate in this region.