We prospectively studied diarrhoea incidence among 1655 children < 5 years of age in northern Vietnam for 1 year using primarily passive surveillance. Standard culture methods were used to detect bacterial pathogens. Overall 2160 cases occurred (1·3 cases/child per year). Peak rates of diarrhoea occurred in children < 12 months old. Rates ranged from 3·3 cases/child per year in children < 1 year old, to 0·7 cases/child per year in 4-year-olds. Campylobacter, shigella and enterotoxigenic Escherichia coli were most commonly isolated. Rates detected by active surveillance were about twice those detected passively. S. flexneri was the most common shigella serogroup (65%). S. flexneri serotypes 6, 4, 1 and Y were most common, but 40% were untypable using commercial antisera. The data illustrate important regional differences in pathogen prevalence and shigella serotype distribution. Shigella vaccine development strategies, commonly targeting S. flexneri 2a, S. sonnei and S. dysenteriae 1, will have little impact on diarrhoea rates in Vietnam.

The Republic of Ireland has the highest incidence of meningococcal disease in Europe with 40% of all cases occurring in children under the age of 5 years. Attending day-care increases the risk of certain infections, including Haemophilus influenzae type b (Hib) meningitis. The risk of meningococcal disease associated with day-care is not known. We conducted a case-control study among pre-school children with 130 laboratory-confirmed cases and 390 controls, matched on age, gender and place of residence, to determine if day-care attendance was a risk factor for meningococcal disease. Multivariate analysis showed that day-care attenders had a lower risk of disease than non-attenders (OR 0·3, 95% CI 0·1–0·7) whereas the number of adults in a household, and household crowding were independent risk factors for disease. Asymptomatic carriers of Neisseria meningitidis are the main source of transmission and these carriers are usually adults. Regular day-care attendance may reduce this risk by removing children from close and prolonged contact with adults.

The incidence of pertussis requiring hospitalization in children younger than 16 years was estimated by the use of an active surveillance-system. Of special interest were differences between West and East Germany following different vaccination strategies before reunification. In 1997 and 1998, 754 pertussis cases required a total of 11151 hospital inpatient days. The incidence of hospitalized pertussis was 2·68/100000 person years and this was significantly higher in East than in West Germany. In East Germany an unusually high percentage of hospitalized cases was found in children aged 6–15 years (45% versus 13% in West Germany). The difference between the regions may be due either to a different perception of the disease or to an increased immunity induced by prior disease or vaccination. In East Germany, pertussis was rare until reunification but it has increased significantly since then. Older children may thus represent a population at risk of pertussis having not had previous exposure to pertussis antigens.

We conducted a study in a squatter settlement in Karachi, Pakistan where residents report commonly washing their hands to determine if providing soap, encouraging hand washing, and improving wash-water quality would improve hand cleanliness. We allocated interventions to 75 mothers and collected hand-rinse samples on unannounced visits. In the final model compared with mothers who received no hand-washing intervention, mothers who received soap would be expected to have 65% fewer thermotolerant coliform bacteria on their hands (95% CI 40%, 79%) and mothers who received soap, a safe water storage vessel, hypochlorite for water treatment, and instructions to wash their hands with soap and chlorinated water would be expected to have 74% fewer (95% CI 57%, 84%). The difference between those who received soap alone, and those who received soap plus the safe water vessel was not significant (P = 0·26). Providing soap and promoting hand washing measurably improved mothers' hand cleanliness even when used with contaminated water.

In the Netherlands a strong epidemic outbreak of pertussis took place in 1996–7. Here we investigate the possible causes of the epidemic, using an age-structured epidemic model. Motivated by the observation that during the epidemic the number of cases in vaccinated children had increased considerably compared to the preceding period, we focus on two vaccination related changes. First, we consider the possibility that the potency of the vaccine decreased so that it confers protection for a shorter period of time in newly vaccinated children. Second, we consider the possibility that at a certain point in time the duration of protection after vaccination decreased for all individuals. This may be the case if the pathogen population changed such that the current vaccine confers less protection. A comparison of the observed and simulated age-distribution of infections indicates that the second scenario is more in line with the observed pattern of the 1996–7 epidemic. We discuss the implications of this conclusion for B. pertussis circulation, and for the design of vaccination programmes in the face of a polymorphic B. pertussis population that may adapt itself to vaccination.

We assessed the prevalence of carriage of methicillin-resistant Staphylococcus aureus (MRSA) in anterior nares and stools, and of third-generation cephalosporin resistant enterobacteriaceae and non-fermenting Gram-negative bacilli (RE/RNF) in stools of 748 hospitalized long-stay cirrhotic patients. We also evaluated the consequences of carriage on the epidemiology of hospital-acquired spontaneous bacterial peritonitis, bacteraemia and urinary tract infection (UTI) in these patients. The prevalence of carriage of MRSA and RE/RNF was 16·7% and 14·7% respectively. Whereas RE/RNF carriage did not lead to an increased risk of infection due to RE/RNF, the overall risk of infections caused by MRSA was more than tenfold higher in MRSA carriers. MRSA and RE/RNF carriers had received prior antibiotic therapy to a greater extent than non-carriers (P < 0·001) and MRSA carriers had received prior norfloxacin prophylaxis to a greater extent than the two other groups (P < 0·02). The mortality rate during hospital stay was higher in MRSA and RE/RNF carriers than in non-carriers (P < 0·001). Pugh score (P < 0·0001), age (P < 0·0001), MRSA carriage (P = 0·0018) and bacteraemia (P = 0·0017) were associated independently with mortality. MRSA carriage in hospitalized cirrhotic patients leads to the emergence of infections due to this strain, mainly SBP and bacteraemia. Prior antibiotic therapy and norfloxacin prophylaxis increase the risk of carriage of MRSA.

We analysed a strain collection representative of the overall Neisseria meningitidis population circulating in an open community (46000 inhabitants, Spain) during an endemic period (30 isolates from patients and 191 from throat cultures of healthy individuals) by both phenotypic and molecular techniques. Almost all patient isolates were assigned to three hyper-virulent lineages (ET-5 complex, ET-37 complex and cluster A4) by both multilocus enzyme electrophoresis (MEE) and pulsed-field gel electrophoresis (PFGE). In contrast, MEE and PFGE assigned 20% and 15% respectively of carrier isolates to the hyper-virulent clones (4% for both methods together). There was also a higher correlation between PFGE and phenotypes associated with virulent clones. These notable differences between the two molecular methods were further observed in more than half the carrier isolates, suggesting that the associations between these strains were distorted by recombination events. However, almost one-third of total endemic strains from symptom-free carriers and almost all patient strains belonged to clones defined by MEE and PFGE, with no known epidemiological connection. These data indicate low transmission and a weak clonal structure for N. meningitidis.

In Victoria between 1990 and 1996, meningococcal infections occurred in 1–2/100000 people each year, with sometimes devastating outcome. In 1997, a typical year, we conducted a case-control study of all cases notified to the State Disease Control Unit, to investigate personal, environmental and lifestyle risk factors. In bivariate analysis many exposures were statistically significantly different (at P = 0·01) in cases and controls. The level of risk, and specific risks, differed between children (under 16) and adults (16 years and over). In multivariate analysis few exposures remained significant (at P = 0·05). However, these included having a smoker amongst close contacts, exposure to construction dust, recent illness, a history of snoring and speech problems, and sharing a bedroom. Besides confirming some previously identified risk factors, this is the first time that snoring and speech problems have been identified as risk factors for meningococcal disease.

The purpose of the study was to describe the results of follow up of human contacts of bovine tuberculosis. The bovine tuberculosis cases occurred on farms in North Staffordshire between 1993 and 1997. A total of 162 people were identified as having close contact with cattle diagnosed as having bovine tuberculosis, or who had drunk unpasteurized milk from a herd with bovine tuberculosis. A retrospective review of chest clinic notes was performed. One hundred and thirty-eight people attended for follow up, and Heaf test results, necessity for chest X-ray and further clinical follow-up are described. No case of human Mycobacterium bovis infection was identified. It is suggested that follow-up of human contacts is limited to those with close contact with herds who have bovine tuberculosis and cattle with visible pulmonary lesions or evidence of udder infection. Children on the farms with affected cattle should also be offered BCG in advance of the routine school's programme.

The immunogenicity and reactogenicity of the tetanus–diphtheria adult type vaccine was compared in two groups: group I (n = 201, 18–30 years old, presumably vaccinated with the DTP vaccine) and group II (n = 147, [ges ] 45 years old, without vaccination antecedents). Before vaccination, the seroprotection levels for tetanus were 90·5% (group I) and 30·6% (group II). These rose to 99·5% and 81·7%, respectively, after administration of one vaccine dose. For diphtheria, prevaccination seroprotection levels were 38·3% (group I) and 19·0% (group II). These rose to 85·8% and 65·7%, respectively, after vaccination. The logistic regression analysis showed an association between antibody titre and age. In group II, 3 doses of Td vaccine were needed to reach titres similar to those achieved in group I with a single dose. Stated reactogenicity was greater in: young subjects, women, those with higher titres of tetanus antibodies and those receiving other vaccines simultaneously. These results confirm the need for vaccination schedules adapted to the characteristics of each population age-group.

One fifth of 527 cases of hepatitis A occurred in self-identified injection drug users during a community-wide epidemic in Spokane County (Washington) in 1997–8. We hypothesized that an immunization campaign targeted at illicit drug users could control the epidemic. Starting in May 1998, hepatitis A vaccine was provided to individuals in jails and other sites frequented by illicit drug users. Volunteers at vaccination sites were surveyed about risk. Serial convenience samples of jail inmates who denied previous vaccination were anonymously tested for hepatitis A virus (HAV) immunoglobulin G (IgG). From May to December 1998, 2765 high-risk individuals were vaccinated against hepatitis A. The proportion of HAV IgG seropositive inmates increased from 30% to more than 50%. Our findings suggest that vaccination along with naturally occurring infection increased the rate of hepatitis A immunity among illicit drug users during the final months of the epidemic. This supports the hypothesis that targeted immunization of high risk groups may shorten the natural history of a community-wide epidemic.

Individuals developing an HIV seroconversion illness may experience rapid disease progression. Information on seroconversion illness is rarely collected in most cohort studies; thus the aim of this study was to assess the value of the HIV test interval (the time between last negative and first positive HIV tests) as a proxy for seroconversion illness. Among 8229 seroconverters, test intervals ranged from 0–5282 days, and varied by gender, risk group, age and calendar year of seroconversion. Those with intervals [les ] 31 days had an increased hazard of AIDS (RH 1·42, P = 0·07), which was reduced slightly after adjusting for baseline factors, calendar year of follow-up, treatment and the declining CD4 count, but there was no effect on survival. Thus, it appears that if information on acute seroconversion illness is not available, then analyses of progression to AIDS in seroconverter studies could use a short test interval as a proxy measure.

We examined the use of polymerase chain reaction (PCR) in investigating suspected cases of meningococcal infection in Birmingham. Data held by Birmingham Health Authority were interrogated to determine cases of suspected or confirmed meningococcal infection for a 3-year period from April 1996. The microbiology departments of five local hospitals completed a standard proforma about the microbiological investigation of cases and included details of patient age, clinical presentation and method of confirmation of the clinical diagnosis. Of 273 cases, 123 had PCR performed on either cerebrospinal fluid and/or blood. Groups more likely to have a PCR done were those presenting with septicaemia alone, and those in the 5–14 year age group. In 33 cases, PCR was the only positive microbiological result. Over the study period there was increasing but variable use of PCR in the investigation of meningococcal infection and PCR increased the yield of confirmed cases.

This descriptive study investigated an outbreak of hepatitis A virus (HAV) infection among injecting drug users (IDUs) and their contacts. Twenty-seven cases of acute HAV infection were identified in a 5-month period. Connections with the local injecting drug using (IDU) population were established for 25 of the cases of whom 14 admitted to injecting drug use. HAV RNA genotyping revealed two HAV variants, closely related to variants found in Scandinavian IDUs and in South East Asia. The study demonstrates that once HAV enters the IDU population extensive outbreaks are possible. We recommend that all IDUs should be tested for HAV and hepatitis B virus (HBV) infections and offered combined hepatitis A and B vaccines if non-immune.

Legionnaires' disease is an uncommon but important cause of life-threatening community-acquired or nosocomial pneumonia. The urinary antigen enzyme immunoassay test, used in Victoria since 1995, now accounts for the majority of initial laboratory notifications (81% in 1999). We review the impact of the test on the disease epidemiology and the public health investigative process. We focus on the major subgroup of cases due to Legionella pneumophila serogroup 1, comparing delays until notification and mortality for urinary antigen detected cases with culture detected cases. The urinary antigen test facilitates a 5-day reduction for the delay between onset of illness and notification. We observed that there was minimal clinical heterogeneity of urinary antigen detected cases whether they were subsequently culture confirmed or not. We encourage clinician use of the urinary antigen test in cases of community-acquired pneumonia where Legionnaires' disease is a possible diagnosis, in conjunction with culture of clinical specimens.

In order to assess the feasibility of environmental poliovirus surveillance, known amounts of poliovirus type 1, strain Sabin, were flushed into the sewage network of Helsinki. Grab specimens collected at a remote downstream location and concentrated about a 100-fold revealed infectious poliovirus on four successive days in all three separate experiments. As for concentration, a simple two-phase separation method was found to be at least as useful as a several-fold more resource-demanding polyethylene glycol (PEG) precipitation method. Recovery of the introduced virus was remarkably high (more than 10%). Using the current system, it might be possible to detect poliovirus circulation in a population of 700 000 people by examining a single 400 ml sewage specimen, if 1 out of 10000 inhabitants were excreting the virus. It is concluded that environmental surveillance is a sensitive approach to monitor silent poliovirus circulation in populations served by a sewage network.

The objective of this study is to estimate the measles reproduction ratio for eight Western European vaccination programmes. Because many plausible age-structured transmission patterns result in a similar description of the observations, it is not possible to estimate a unique value of the reproduction ratio. A method is developed to estimate bounds and confidence intervals for plausible values of the reproduction ratios using maximum likelihood methods. Lower and upper bounds for plausible values of the basic reproduction ratio are estimated to be 7·17 (95% CI 7·14–7·20) and 45·41 (95% CI 9·77–49·57), corresponding to lower and upper bounds on critical vaccine coverage of 86·6% and 98·1%. Of the eight evaluated vaccination programmes, four have vaccine coverage below the lower bound and allow measles to persist, and four have vaccine coverage at the upper bound and may eventually eliminate measles.

This study aimed to assess the seroprevalence and risk factors for hepatitis B virus (HBV), hepatitis C virus (HCV), and HIV-1 infections among injecting drug users (IDU) in New Mexico. Serological and behavioural surveys were conducted in conjunction with street-based outreach, education and HIV counselling and testing. High rates of antibody positivity for HCV (82·2%) and HBV (61·1%), and a low rate for HIV (0·5%) were found. In multivariate analyses, both HBV and HCV infection were positively associated with increasing age, increasing years of injection and heroin use. Receipt of a tattoo in prison/jail was associated with HBV (odds ratio = 2·3, 95% confidence interval 1·4, 3·8) and HCV (OR = 3·4, 95% CI = 1·6, 7·5) infections. Prevention of bloodborne pathogens among IDUs should focus on young users, early in their drug use experience. Studies examining the relationship between tattooing and HBV and HCV infection are needed as are efforts to promote sterile tattooing, in prisons and elsewhere.

Two epidemics of acute, watery diarrhoea in villages in North Arcot district, India, were investigated. The attack rates were 10·03 and 15·53 per 100 population, the median duration was 5 days and enteric pathogens were present in 56·8% and 60·3% of specimens from the two villages, but no predominant pathogen was identified. Examination of stools from a 20% age-stratified random sample of the population of one of the villages after the epidemic found 22·9% of asymptomatic subjects excreted bacterial enteric pathogens. Despite the high background of enteric pathogen carriage, the isolation rates for shigellae, enteropathogenic Escherichia coli and Shiga-toxin producing E. coli were significantly higher (P < 0·001, P < 0·02, P < 0·05) during the epidemic. The epidemics may have been caused by faecal contamination of well water following rain. Point-of-use techniques for water disinfection may be most effective for preventing such outbreaks, but further research into the development of appropriate technology is required.

To evaluate the impact of the 1998 Australian Measles Control Campaign on immunity to measles and rubella, 4400 opportunistically-collected sera, submitted to diagnostic laboratories across Australia from subjects aged 1–49 years, and 3000 from subjects aged 1–18-years, were tested before and after the campaign, respectively. The proportion of individuals aged 1–18 years who were immune to measles rose from 85% before, to 90% after, the campaign (P < 0·001). The greatest increase was in preschool (7%, P < 0·001) and primary school (10%, P < 0·001) children, who were actively targeted by the campaign. Rubella immunity in 1–18 year-olds rose from 83% to 91% (P < 0·001), again with significant increases in preschool (4%, P = 0·002) and primary school (16%, P < 0·001) children. 94% of individuals aged 19–49 years were immune to rubella. These serosurveys confirm other evidence of the effectiveness of the Australian Measles Control Campaign and demonstrate the value of serosurveillance using opportunistically collected sera.