We tested symptom-based surveillance during the G8 conference in 2000 as a means of detecting outbreaks, including bio-terrorism attacks, promptly. Five categories of symptoms (skin and haemorrhagic, respiratory, gastrointestinal, neurological and unexplained) were adopted for the case definition of the surveillance. The surveillance began 1 week before the conference, and continued until 1 week after the conference ended. We could not detect any outbreaks during this surveillance. Compared to the existing diagnosis-based surveillance system, symptom-based surveillance has the advantages of timeliness and simplicity. However, poor specificity and difficulties in determining epidemic threshold were important limitations of this system. To increase the specificity of surveillance, it is essential to incorporate rapid laboratory diagnoses into the system.

Campylobacter jejuni isolates of human, canine, feline, bovine and poultry origin were investigated for their genomic diversity using O-antigen typing (n = 271), SmaI (n = 158) and XhoI (n = 158) macrorestriction analysis and ERIC–PCR (n = 107). The O-antigens O[ratio ]1/44, O[ratio ]2, O[ratio ]4 complex, O[ratio ]37, O[ratio ]40 were identified and 53.7% of the human and 56.1% of the animal strains were typable with the available antisera. Two ERIC–PCR pattern groups were generated representing human and animal strains as well as those exclusively of animal origin. XhoI macrorestriction analysis also distinguished ‘human’ and ‘non-human’ strain clusters, but by SmaI restriction mainly serotype-associated clusters were found. In conclusion, genomic differences may occur between ‘human’ and ‘non-human’ strains and this may reflect their potential to overcome the barrier from animals to humans.

Shigella flexneri infections are one of the main causes of acute diarrhoea in Cuba. Twenty strains isolated from sporadic cases in nine different Cuban provinces were characterized. Serotyping, antibiotic-resistance typing, plasmid-typing and AFLP-typing were used to determine their suitability for use in epidemiological studies of S. flexneri. The predominant serotypes were serotype 6 (35%) and serotype 2 (35%). Eleven different plasmid profiles were detected (Diversity Index=0·92). AFLP-typing discriminated 12 different patterns (DI=0·95), these patterns were not coincident with plasmid-typing patterns. Both techniques combined distinguished 14 patterns among the 20 studied strains (DI=0·99). There was no consistent relationship between plasmid-typing and AFLP-typing patterns or antibiotic-resistance typing patterns. Ninety-five percent of S. flexneri strains were multiresistant.

In March 1999, a large community outbreak of Escherichia coli O157 infection occurred in North Cumbria. A total of 114 individuals were reported to the Outbreak Control Team (OCT); 88 had laboratory confirmed E. coli O157. Twenty-eight (32%) of the confirmed cases were admitted to hospital, including three children (3·4%) with haemolytic uraemic syndrome. There were no deaths. A case-control study found that illness was strongly associated with drinking pasteurized milk from a local farm (P=<0·0001) on single variable analysis. Microbiological investigations at the farm revealed E. coli O157 phage type (PT) 21/28 VT 2 which was indistinguishable from the human isolates by pulsed field gel electrophoresis. At the time of occurrence this was the largest E. coli O157 outbreak in England and Wales and the first E. coli O157 PT 21/28 VT 2 outbreak associated with pasteurized milk. This outbreak highlights lessons to be learnt regarding on-farm pasteurization.

Enhanced surveillance of meningococcal disease (ESMD) was implemented nationally across ten regions of England, Wales and Northern Ireland from 1 January 1999. It aims to deliver more sensitive surveillance than laboratory reporting by including clinically diagnosed but laboratory unconfirmed cases. Consultants in Communicable Disease Control (CsCDC) report all clinically diagnosed cases of meningococcal disease (MD) to the Regional Epidemiologist in the relevant regional unit of the Public Health Laboratory Service (PHLS) Communicable Disease Surveillance Centre (CDSC). These reports are reconciled with laboratory data from the PHLS Meningococcal Reference Unit and then forwarded to the national CDSC where further reconciliation with laboratory data takes place. In addition, CsCDC are asked to report any clusters of MD that occur. Between 1 January 1999 and 30 June 2001, 12074 cases of MD were ascertained through ESMD. The majority (57%) were laboratory confirmed. The estimated incidence of MD fell between 1999 and 2001 from 9.2 to 8.0 per 100000 population. Of laboratory confirmed cases, the number of cases of serogroups B and W135 increased and of serogroup C and of ungrouped meningococcal infection decreased. Variation between regions was considerable and deserves further investigation. Of 11522 cases with a reported clinical diagnosis, 53.6% were diagnosed as septicaemia, 32.6% as meningitis, 12.5% as both septicaemia and meningitis, and 1.3% had other invasive MD. Between 1 January 1999 and 30 June 2001 698 deaths were reported, an overall case fatality rate (CFR) of 5.8%; 567 deaths were in confirmed cases and 131 probable (CFR 8.2% and 2.5%, respectively). CFR was higher in serogroup C (13.5%) than B (5.8%). No peak in serogroup C meningococcal infection occurred in the winter of 2000/1 and no clusters of serogroup C meningococcal infection were reported in the first half of 2001. ESMD provides information about the epidemiology of MD that is more complete than statutory notification and laboratory surveillance and is useful for evaluating the impact of the meningococcal serogroup C vaccination programme and of the other non-vaccine preventable serogroups.

To describe the epidemiology of invasive group A streptococcal (iGAS) infections in the San Francisco Bay Area, population-based active surveillance for laboratory-confirmed iGAS was conducted by the California Emerging Infections Program in three California counties. From January 1989 to December 1999, 1415 cases of iGAS were identified. Mean iGAS incidence was 4·06/100 000 person-years and case fatality ratio was 13%, with no linear trends over time. Incidence was lowest in adolescents, was higher in men than women (4·4 vs. 3·2/100 000 person-years), and was higher in African–Americans (6·7) than in non-Hispanic (4·1) or Hispanic (3·4) Whites, Asians (2·2) or Native Americans (1·7/100 000 person-years). Injecting drug use was the riskiest underlying condition and was associated with the highest attributable risk. Cases were associated with several underlying conditions, but 23% occurred in previously healthy persons. From 1989–1999, iGAS infections in the San Francisco Bay Area became neither more common nor more deadly.

We aimed to provide a quantitative description of decay in pertussis antibody levels to aid in finding a serological estimate of the incidence of pertussis. The serum IgG response against pertussis toxin was studied in a group of clinically diagnosed patients. Individual records consisted of repeated serum IgG measurements at irregular intervals for up to 10 years post diagnosis. These data were analysed with a nonlinear regression model taking into account censoring at upper and lower threshold levels, measurement errors, and individual variation in the shape and magnitude of the immune response. There was considerable variation between individual responses, both in strength (amplitude) and duration (shape). The inverse model relating IgG levels to time from infection (diagnosis) can be applied to cross-sectional IgG data to generate distributions of times from infection, which may be used to calculate infection rates and their variation, in populations sampled for cross-sectional IgG data.

Streptococcus pneumoniae is the leading bacterial cause of life-threatening infections in infants. Although antibiotic resistance affects management of pneumococcal infections, few data on patterns of resistance are available for India. We examined nasopharyngeal carriage of antibiotic-resistant pneumococci in 464 South Indian infants between 2 and 6 months. Newly acquired serotypes were screened for susceptibility to cotrimoxazole, erythromycin and penicillin using disk diffusion. Cumulative prevalence of pneumococcal carriage rose from 53·9% at 2 months to 70·2% at 6 months. The prevalence of strains that were not susceptible to penicillin, co-trimoxazole and erythromycin was 3·4, 81·1 and 37·2%, respectively. Carriage of erythromycin non-susceptible strains declined significantly between ages 4 months and 6 months (44·1 vs. 10·7%). More than 87% of the isolates screened were non-susceptible to [ges ]1 antibiotic. Serogroups/types that were most frequently non-susceptible to 1 or more antibiotics were 6, 9, 14, 19 and 23. Less than 1% of the isolates were multi-drug resistant. Widespread use of antibiotics in South India has resulted in S. pneumoniae becoming non-susceptible to some commonly used antibiotics. Monitoring trends in antibiotic susceptibility and making antibiotics available only through prescription from a health care worker may slow the spread of resistant pneumococci and improve management of pneumococcal infections in South India.

We used microbiology and pharmacy data from health-maintenance organizations to determine whether antibiotic use by a household member increases the risk of penicillin-non-susceptible pneumococcal disease. Though it has been well established that an individual's antibiotic use increases one's risk of antibiotic-resistant infection, it is unclear whether the risk is increased if a member of one's household is exposed to antibiotics. We therefore conducted a case-control study of patients enrolled in health maintenance organizations in Western Washington and Northern California. Cases were defined as individuals with penicillin-non-susceptible pneumococcal infection; controls were individuals with penicillin-susceptible pneumococcal infection. Socioeconomic variables were obtained by linking addresses with 1997 census block group data. One-hundred and thirty-four cases were compared with 798 controls. Individual antibiotic use prior to diagnosis increased the odds of penicillin non-susceptibility, with the strongest effect seen for β-lactam use within 2 months (OR 1·8, 95% CI 1·2, 2·8). When household antibiotic use by persons other than the patient were considered, at 4 months prior to diagnosis there was a trend towards an association between penicillin non-susceptibility and β-lactam antibiotic use, and a possible association in a small subgroup of patients with eye and ear isolates. However, no significant overall pattern of association was seen. We conclude that though antibiotic use of any kind within 2 months prior to diagnosis is associated with an increased risk of penicillin-non-susceptible pneumococcal disease, there is no significant overall pattern of association between household antibiotic use and penicillin-non-susceptible pneumococcal infection.

This study examines the incidence of antibiotic-associated diarrhoea (AAD) in children in the community setting. Parents of 965 children aged 4–6 years and resident in rural/semi-rural South Australia completed a questionnaire on socio-demographic factors, and a 6-week daily diary detailing symptoms of gastroenteritis, antibiotic use, respiratory illness, and contact with someone with gastroenteritis. The incidence of AAD was 32·3%, falling to 23·5% when episodes associated with a respiratory illness were excluded. Respiratory illness in the previous 3 days (OR 6·76, 95% CI 4·87, 9·38), and contact with someone with gastroenteritis in the previous 14 days (OR 1·8, 95% CI 1·48, 2·19), were both associated with gastroenteritis. After adjusting for these, only the first day of antibiotic use was associated with gastroenteritis (OR 3·8, 95% CI 1·8, 8·06). Potential confounding factors, in particular inter-current respiratory illness, need to be considered when examining AAD.

This study characterizes possible confounders that might make cohorts vaccinated and unvaccinated against influenza and pneumococcal infection different at baseline, with the hypothesis that the two cohorts are comparable. The similarity between health and demographic data was analysed by a randomized, multivariant study addressed to 10 000 persons aged 65 years and older in Stockholm County and was carried out in the form of a postal inquiry during the period December 2000 to May 2001. The study-population response rate was 78%. Of these, 66% (5120 persons) had been given at least one influenza vaccination during the 3-year study period (1998–2000), 50% (3780) had received one pneumococcal vaccination and 78% had received both vaccines during the period. The vaccination rate was lower in the age group 65–69 years (60%), compared with elderly cohorts aged over 70 years (67–72%, P<0·001). Elderly persons living in nursing homes or institutions had higher vaccination rates than persons living in their own households (72 vs. 67%). Persons with underlying chronic diseases had higher vaccination rates (71%, P<0·001) than those without underlying chronic diseases. Vaccine recipients had fewer days in hospital, compared with non-recipients. Unvaccinated persons with myocardial disease had nine times more days in hospital than vaccinated persons with myocardial disease. Vaccination against pneumococcal infection had an additional effect with influenza vaccination in reducing hospitalization for chronic lung diseases; influenza vaccination alone did not have this effect. In conclusion, the influenza and pneumococcal-vaccine recipients were older and had significantly more chronic lung and heart conditions than the unvaccinated cohort.

Since new vaccines and anti-viral drugs for influenza have become available, collation of actual and country-specific epidemiological data is essential. Since respiratory syncytial virus (RSV) is a well known paediatric airway pathogen and some epidemiological data exist already, a comparison between influenza and RSV seems warranted. From July 1996 to June 2001 the naso-pharyngeal aspirates (NPA) of children from birth to 16 years of age, admitted to one of the two paediatric hospitals in Kiel, Germany, were investigated by a nine-valent multiplex reverse transcriptase PCR assay. NPA were investigated in 60·8% of 3469 children admitted with an acute respiratory tract infection. Community-acquired or nosocomial infections (in parentheses) due to influenza A were diagnosed in 122 (10) children, due to influenza B in 14 (2) and due to RSV in 325 (24) cases. Patients with influenza A (median 752 days) and influenza B (median 966 days) were older than patients with RSV (median 168 days). The spectrum of disease presentation was broader in influenza than in RSV. In each winter, admissions with influenza were less common than those with RSV. Influenza B only occurred in 2 of the 5 years. The cumulative, population-based incidences per 100 000 children 0–16 (0–5, >5–16) years of age were 53 (123, 22) for influenza A, 16 (30, 9) for influenza B and 165 (453, 4) for RSV. Cardiac conditions and asthma were the major risk factors for admission to hospital with influenza A (RR 9·8, 4·1) and RSV (8·5, 2·1) infections. Underlying conditions were most common in influenza B. Low gestational age doubled the risk for admission to hospital with influenza A infection, but did not show a dose–effect relationship as in RSV. The burden of influenza-positive hospitalizations was about one third that of RSV. The incidence was similar to reports from the United States. Targeting children with underlying conditions, especially cardiac conditions and asthma in the German immunization programme is appropriate, as long as no policy for vaccination of the general paediatric population exists.

In 1996, a seroepidemiological study was undertaken in Spain, with the main aim of estimating the population's immunity against poliomyelitis, tetanus and diphtheria. A population-based cross-sectional study was conducted, covering the population aged 2–39 years. The sample was stratified by age and rural–urban environment, and informed consent obtained to take blood specimens from subjects attending phlebotomy centres. The study included 3932 persons and the prevalence of antibodies against all three types of poliovirus exceeded 94% across all age groups. From a high of 96% in subjects under the age of 15 years, immunity against diphtheria steadily declined to a low of 32·3% in subjects aged 30–39 years. Similarly, tetanus antitoxin concentrations indicating basic protection were present in 98–99% of the under-14 years age group; thereafter, immunity declined, until reaching 54·6% in the 30–39 years age group.

A serosurvey for measles, mumps and rubella was conducted in Italy; incidence based on statutory notifications over the last three decades was also calculated. In Italy the diseases followed an endemic–epidemic pattern, with an incidence peak every 2–4 years, and had a limited reduction of incidence attributable to childhood immunization. Lower notification rates were observed in the Southern regions. This is possibly related to greater under notification in the South and is confirmed by our seroprevalence data. Incidence of measles and rubella and proportion of cases among young adults increased significantly in the three decades considered, but not for mumps. Serological data confirmed that these infections are still very frequent in Italy, without significant geographic variation in the country. In the age groups 2–4 and 5–9 years the percentage of individuals still susceptible to each virus was higher than 30%. The proportion of susceptible subjects older than 15 years was similar for the three infections (6·1, 11·7 and 8·8% for measles, mumps and rubella, respectively). The low vaccine coverage for rubella and measles in Italy has so far only partially affected the occurrence of the diseases. No impact of mumps vaccination is visible. The average number of deaths, for each disease, has decreased during the three study periods. Today the priority in Italy is to halt the progressive increase of the mean age of acquisition of the three infections, to eliminate differences in coverage among regions and to conform to European standards. This will be achieved through a combination of increasing MMR vaccine coverage before 2 years of age, implementing vaccination campaigns for low seroprevalence age groups, and/or introducing a second dose of MMR, depending on the level of current MMR coverage.

A mumps outbreak occurred in a group of vaccinated children aged 3–4 years in San Sebastián (Gipuzkoa, Basque Country, Spain) in 2000 during the same period as a revaccination campaign against measles–mumps–rubella (MMR) was performed. The clinical cases were confirmed by viral culture, detection of viral RNA and/or specific IgM. Eighty-eight percent of the children had been vaccinated with the Rubini strain and the remainder with the Jeryl–Lynn strain. The attack rate was 47·9% (35 cases in 73 school-attending children of this age). The outbreak was caused by an H genotype strain of mumps virus which was circulating at the same time as a D genotype strain that caused sporadic cases. By sequencing the small hydrophobic (SH) gene, the strains of the clinical cases were identified as wild-type mumps virus with heterologous genotypes in comparison to the vaccine strains used in our area.

We followed students in eight elementary schools for rubella antibody from 1993 to 1996 (602 pairs) and 1996–9 (588 pairs) in Gyeonggi Province, Korea. We tested rubella IgG and administered rubella vaccine to the children with the titres <10 IU/ml. The loss rates of rubella IgG during the follow-up periods were 14·3 and 15·8%, respectively. Among vaccinated groups, the loss rate was 18·8%, which was significantly higher than 13·8% of the mixture of natural and vaccine-induced immunity groups. The group that had the lower preceding antibody titre had a higher loss rate of 24·8% compared to 7·2% for the group whose titre was 40 IU/ml or above. In a multivariate analysis, age and gender were not related to antibody loss rate. Under this higher rubella antibody loss rate, in order to prevent congenital rubella syndrome, the immunization for women at childbearing age appears necessary until rubella can be eliminated or controlled.

In a prospective observational study 4485 patients from 46 clinical centres in 17 European countries were followed between April 1994 and November 1996. Information on AIDS-defining events (ADEs) were collected together with basic demographic data, treatment history and laboratory results. The centres were divided into four geographical regions (north, central, south-west and south-east) so that it was possible to identify any existing regional differences in ADEs. The regional differences that we observed included a higher risk of all forms of Mycobacterium tuberculosis infections (Tb) and wasting disease in the south-west and an increased risk of infections with the Mycobacterium avium complex (MAC) in the north. In Cox multivariable analyses, where north was used as the reference group, we observed hazard ratios of 6·87, 7·77, 2·29 and 0·16 (P<0·05 in all cases) for pulmonary Tb, extrapulmonary Tb, wasting disease and MAC respectively in the south-west. Pneumocystis carinii pneumonia (PCP) was less commonly diagnosed in the central region (RH=0·51, 95% CI 0·32–0·79, P=0·003) and most common in the south-east (RH=1·04, 95% CI 0·71–1·51, P=0·85). Comparisons with a similar ‘AIDS in Europe’ study that concentrated on the early phase of the epidemic reveal that most of the regional differences that were observed in the 1980s still persist in the mid-1990s.

Hepatitis C virus (HCV) among injecting drug users (IDUs) is one of the European Union's (EU) major public health problems. This review examines the current state of knowledge regarding HCV among IDUs in EU countries. Studies published between January 1990 and December 2000, were identified through a computerized search (MEDLINE and EMBASE). Ninety-eight studies have reported prevalence for HCV among groups of IDUs in all EU countries except Luxembourg. The prevalence of anti-HCV ranged from 30 to 98%. Incidence rates ranged from 6·2 to 39·3 per 100 person years. This review provides a comprehensive examination of HCV infection among IDUs in the countries of the EU, and quite clearly demonstrates that the quality and epidemiological relevance of the studies published varies widely. Thus, the reported data may not reflect accurately the current or recent past prevalence of HCV among IDUs in the EU. A strategic approach to the surveillance of HCV among IDUs in the EU, utilizing robust and consistent methods, is required urgently.

The English HCV lookback programme has identified some individuals with transfusion-transmitted HCV infection. The path from the collection of donations from HCV-infected donors to the identification of infected recipients was constructed. The probability of different outcomes at each branch was derived from data collected during this programme. This path of probabilities was then used to produce a complete estimate of the number of recipients infected by blood transfusions (dead and alive at the end of 1995) by re-entry of blood components that fell out of the lookback at various steps prior to recipient testing, and entry of components from HCV-infected donations that were never identified for lookback. Less than 14 000 recipients were estimated to have been infected with HCV during the decade prior to the start of donation testing. Over 60% of these were expected to have died by the end of 1995. Transfusion has infected a large group of individuals. However, this group constitutes a very small, and declining, proportion of all HCV infections in the population.

Ultraviolet radiation (UVR) suppresses many aspects of cell-mediated immunity but it is uncertain whether solar UV exposure alters resistance to human infectious diseases. Varicella-zoster virus (VZV) causes varicella (chickenpox) and can reactivate from latency to cause zoster (shingles). The monthly incidence of chickenpox and zoster in a defined Polish population over 2 years was recorded and ground level solar UV was measured daily. There was a significant seasonality of UVR. Evidence of seasonal variation was found for all zoster cases and for zoster in males, with the lowest number of cases in the winter. The number of zoster cases with lesions occurring on exposed body sites (the face) demonstrated highly significant seasonality with a peak in July/August. Seasonal models for UVR and zoster cases showed similar temporal patterns. By contrast, for varicella, the maximum number of cases was found in March and the minimum in August/September, probably explained by the respiratory spread of VZV. It is tempting to speculate that the increase in solar UVR in the summer could induce suppression of cellular immunity, thus contributing to the corresponding rise in the incidence of zoster.