The emergence of Vibrio cholerae O139 in 1992 and reports of an increasing number of other non-O1 serogroups being associated with diarrhoea, stimulated us to characterize V. cholerae non-O1 non-O139 strains received at the National Institute of Infectious Diseases, Japan for serotyping. Ribotyping with the restriction enzyme BglI of 103 epidemiological unrelated mainly clinical strains representing 10 O-serotypes yielded 67 different typing patterns. Ribotype similarity within each serotype was compared by using the Dice coefficient (Sd) and different levels of homogeneity were observed (serotypes O5, O41 and O17, Sd between 82 and 90%; serotypes O13 and O141 Sd of 72; and O2, O6, O7, O11, O24 Sd of 62–66%). By cluster analysis, the strains were divided into several clusters of low similarity suggesting a high level of genetic diversity. A low degree of similarity between serotypes and ribotypes was found as strains within a specific serotypes often did not cluster but clustered with strains from other serotypes. However, epidemiological unrelated O5 strains showed identical or closely related ribotypes suggesting that these strains have undergone few genetic changes and may correspond to a clonal line. Surprisingly, 10 of 16 O141 strains studied contained a cholera toxin (CT) gene, including 7 strains recovered from stool and water samples in the United States. This is to our knowledge the first report of CT-positive clinical O141 strains. The closely related ribotypes shown by eight CT-positive strains is disturbing and suggest that these strains may be of a clonal origin and have the potential to cause cholera-like disease. Despite the low degree of correlation found between ribotypes and serotypes, both methods appears to be valuable techniques in studying the epidemiology of emerging serotypes of V. cholerae.

Experiments with 2 wild type isolates of Salmonella enterica serotype Enteritidis PT4, which differed in RpoS expression, tolerance to certain hostile environments and pathogenicity, found that changes in in vitro acid, heat, or peroxide tolerance had no effect on the ability of the isolates to multiply in the spleens of C57/BL7/J mice infected orally. Thus, with the pathogenic RpoS-positive isolate, the infectivity of log phase chilled cells, which are profoundly acid-sensitive, was the same as that of non-chilled stationary phase cells which are acid-tolerant. Similarity the infectivity of the RpoS-negative, sensitive isolate, was not enhanced by increases in any tolerance. The ability to survive on surfaces, like infectivity, was also largely unaffected by either growth phase or cold exposure. These two attributes may thus be related and, given that the pathogenic PT4 isolate is capable of prolonged survival and the non-pathogenic isolate survives poorly, survival could serve as a potential marker of pathogenicity. Although the pathogenicity of the two isolates was very different, they showed an almost identical increase in acid tolerance following culture at pH 4·0 for up to 60 min.

The prevalence of campylobacter gastroenteritis has been estimated by bacterial isolation using selective culture. However, there is evidence that certain species and strains are not recovered on selective agars. We have therefore compared direct PCR assays of faecal samples with campylobacter culture, and explored the potential of PCR for simultaneous detection and identification to the species level. Two hundred unselected faecal samples from cases of acute gastroenteritis were cultured on modified charcoal cefoperazone deoxycholate agar and subjected to DNA extraction and PCR assay. Culture on CCDA indicated that 16 of the 200 samples contained ‘Campylobacter spp.’. By contrast, PCR assays detected campylobacters in 19 of the 200 samples, including 15 of the culture-positive samples, and further identified them as: C. jejuni (16), C. coli (2) and C. hyointestinalis (1). These results show that PCR offers a different perspective on the incidence and identity of campylobacters in human gastroenteritis.

M protein gene typing was used to analyse Streptococcus pyogenes clinical isolates collected between 1983 and 1995 in an area of central Italy from patients presenting different types of infections; the same isolates were also characterized by means of DNA fingerprinting.

M type 1 was the most common (50% of study strains), followed by M types 4, 12 and 6. The proportion of M type 12 decreased with time, whereas M type 1 increased, in agreement with data obtained in many different areas. Most invasive strains belonged to types M1 (30%) and M12 (30%); on the other hand, the M1 type did frequently occur also among non-invasive isolates. DNA fingerprinting showed a correlation between M types and DNA patterns. This report provides epidemiological information from a geographic area not sampled recently, and further shows the usefulness of the M genotyping technique, which offers potential advantages over conventional serological typing methods.

The relative frequency of salmonella strains isolated from hospitalized and non-hospitalized patients in Southern Israel changed during the period, 1994–6. Salmonella enterica serotype Typhimurium definitive phage-type 104 (DT104) appeared in Israel in 1994 and became the most prevalent strain in 1996. An outbreak of enteritis due to Salmonella enterica serotype Agona occurred in Israel, in October 1994 and lasted for 4 months. The relative frequency of Salmonella enterica serotype Enteritidis remained almost constant during these years, with seasonal fluctuations only.

The importance of the increase in the prevalence of Typhimurium DT104 has been the epidemic spread of a multiresistant strain of R-type ACT (A, ampicillin; C, chloramphenicol; T, tetracycline) belonging to this phage-type. Since 1995 the frequency of Typhimurium DT104 isolates that possess, in addition to the above R-type, a chromosomally encoded resistance to the quinolone drug, nalidixic acid, increased tenfold. In 1996, 27% of the Typhimurium DT104 isolates were of R-type ACTN. S. Enteritidis exhibited over 95% susceptibility to at least eight of the most commonly used antibiotic drugs, and none of the isolates was resistant to quinolone or fluoroquinoline.

Three cohorts of Danish male military recruits (n=1069) were studied for pharyngeal meningococcal carriage during 3 months at different seasons: 39–47% of entrants were meningococcal carriers and the carriage rate remained constant over time and season. However, individual changes in the carrier state occurred frequently, and after 3 months 34% had changed carrier state on one or more occasions. Initially, a loss of carriage predominated; on the other hand almost 20% of non-carriers had acquisition of meningococci within the first month. The serological phenotypes of the 670 carrier strains were compared with those of 261 invasive strains recovered concurrently from patients with meningococcal disease country-wide. Both carrier strains and invasive strains were phenotypically heterogeneous. Almost 60% of the invasive strains belonged to three phenotypes: B[ratio ]15[ratio ]P1.7, 16, C[ratio ]2a[ratio ]P1.2, 5 and C[ratio ]2b[ratio ]P1.2, 5. In contrast, these phenotypes only amounted to 3·2% of the carrier strains, among which no phenotype was found with a prevalence above 4·9%. However, 30% of the carrier strains had serological phenotypes identical to those of 80% of the invasive strains. Our results indicated that the transmission rate of potential pathogenic carrier strains did not differ from that of other carrier strains.

A total of 259 clinical isolates of nonrepetitive non-typhi salmonellae (NTS) were examined for antibiotic resistance patterns and plasmid content. The antibiotics used were amoxicillin-clavulanic acid (AMC), ampicillin (AM), aztreonam (ATM), carbenicillin (CB), cefixime (CFM), cefotaxime (CTX), cefoxitin (FOX), ceftazidime (CAZ), ceftriaxone (CRO), chloramphenicol (C), ciprofloxacin (CIP), gentamicin (GM), imipenem (IPM), ofloxacin (OFX), tetracycline (TE), trimethoprim-sulfomethoxazole (SXT). Multi-drug resistant (MDR) strains comprised 19·3% of the total isolates (50/259) and almost all were S. typhimurium (49/50). Fifteen different patterns of resistance was observed, AM/CB/C/AMC/TE and AM/CB/C/AMC/SXT/GM/CTX/CRO/CAZ/CFM/ATM being the most frequent patterns. Twenty-eight out of 50 multiresistant isolates were found to contain at least one plasmid (mean five) and the size of the plasmids ranged between 1·7 and 158 kb. Plasmid profiles of multi-resistant NTS strains were heterogenous as 21 different profiles were detected in a total of 28 plasmid-bearing isolates. No direct correlation was established between antibiotic resistance patterns and plasmid profiles.

The net and basic reproduction numbers are among the most widely-applied concepts in infectious disease epidemiology. A net reproduction number (the average number of secondary infectious cases resulting from each case in a given population) of above 1 is conventionally associated with an increase in incidence; the basic reproduction number (defined analogously for a ‘totally susceptible’ population) provides a standard measure of the ‘transmission potential’ of an infection. Using a model of the epidemiology of tuberculosis in England and Wales since 1900, we demonstrate that these measures are difficult to apply if disease can follow reinfection, and that they lose their conventional interpretations if important epidemiological parameters, such as the rate of contact between individuals, change over the time interval between successive cases in a chain of transmission (the serial interval).

The net reproduction number for tuberculosis in England and Wales appears to have been approximately 1 from 1900 until 1950, despite concurrent declines in morbidity and mortality rates, and it declined rapidly in the second half of this century. The basic reproduction number declined from about 3 in 1900, reached 2 by 1950, and first fell below 1 in about 1960. Reductions in effective contact between individuals over this period, measured in terms of the average number of individuals to whom each case could transmit the infection, meant that the conventional basic reproduction number measure (which does not consider subsequent changes in epidemiological parameters) for a given year failed to reflect the ‘actual transmission potential’ of the infection. This latter property is better described by a variant of the conventional measure which takes secular trends in contact into account. These results are relevant for the interpretation of trends in any infectious disease for which epidemiological parameters change over time periods comparable to the infectious period, incubation period or serial interval.

To determine risk factors associated with the occurrence of sporadic cases of Salmonella enteritidis infections among children in France, we conducted a matched case-control study. Cases were identified between 1 March and 30 September 1995. One hundred and five pairs of cases and controls matched for age and place of residence were interviewed. In the 1–5 years age group, illness was associated with the consumption of raw eggs or undercooked egg- containing foods (OR 2·4, 95% CI 1·2–4·8). Storing eggs more than 2 weeks after purchase was associated with Salmonella enteritidis infection (OR 3·8, 95% CI 1·4–10·2), particularly during the summer period (OR 6·0, 95% CI 1·3–26·8). Cases were more likely to report a case of diarrhoea in the household 10·3 days before the onset of symptoms, particularly in the age group [les ]1 year (P=0·01). This study confirms the link between eggs and the occurrence of sporadic cases of Salmonella enteritidis among children, highlights the potential role of prolonged egg storage and underlines the role of person-to-person transmission in infants.

Changes in the frequency of serogroup B non serotypable (B[ratio ]NT) meningococci isolated in England and Wales were investigated by T-track fingerprint analysis, DNA nucleotide sequence determination, and serotyping by whole cell ELISA and dot blot assay. Seventy-three per cent of the isolates designated as B[ratio ]NT by the Meningococcal Reference Unit (MRU) dot blot assay during 1993–4, expressed variants of the serotyping antigen, PorB, that were serotype 4 by whole cell ELISA. T-track fingerprint patterns of these and other ‘serotype 4’ isolates revealed five distinct porB alleles which were shown by nucleotide sequence determination to encode different peptide sequences. Differential binding of the ‘serotype 4’ mAbs MN14G21 and 5DC4C8G8 in whole cell ELISA and dot blot assays was the result, (i) of differences in the peptide sequence of predicted surface loop I and (ii) an amino acid deletion in predicted loop VI of the PorB protein.

The epidemiological progression of human salmonellosis in Norway is parallel to trends noted elsewhere in Europe. During the past two decades, the number of reported cases has increased steadily, with a special sharp rise in the early 1980s due to the emergence of Salmonella enteritidis, followed by a levelling off in recent years. However, in contrast to the situation in most other European countries, about 90% of the cases from whom a travel history is available, have acquired their infection abroad. The incidence of indigenous salmonella infections as well as the prevalence of the microorganism in the domestic food chain, are both comparatively low. In 1993–4, a national case-control study of sporadic indigenous salmonella infections was conducted to identify preventable risk factors and guide preventive efforts. Ninety-four case patients and 226 matched population controls were enrolled. The study failed to demonstrate any statistically significant association between salmonellosis and consumption of domestically produced red meat, poultry or eggs. The only factor which remained independently associated with an increased risk in conditional logistic regression analysis, was consumption of poultry purchased abroad during holiday visits to neighbouring countries. A separate analysis of Salmonella typhimurium infections incriminated food from catering establishments and foreign travel among household members, in addition to imported poultry.

Liverpool, a seaport in NW England, suffered severely from lethal infectious diseases in the second half of the 19th century: the population was densely crowded and malnourished and life expectancy was low. Time-series analysis shows that the epidemics of whooping cough (i) had an interepidemic interval of 2·9 years, 1863–85, which lengthened to 3·4 years, 1885–1900 (ii) were strongly coherent with wheat prices (P<0·001) and (iii) also correlated with cycles of seasonal weather conditions. It is suggested from mathematical modelling that the epidemics in this compromised population were maintained (i.e. the system was driven) by an oscillation of malnutrition and by seasonal weather conditions. A model that incorporates both the dynamics of whooping cough and the demographic characteristics of the population is presented. It has been shown to replicate the dynamics of the epidemics and has been used to predict the changes with time of (i) the force of the infection and (ii) the proportion of those infected with whooping cough who died.

The risk of Borrelia burgdorferi infection and the value of antibiotic prophylaxis after tick bite are controversial. In this study, performed in two areas of southwestern Germany, ticks were collected from 730 patients and examined by the polymerase chain reaction (PCR) for B. burgdorferi. To assess whether transmission of B. burgdorferi occurred, the patients were clinically and serologically examined after tick removal and during follow-up examinations. Data from all tick bites gave a total transmission rate of 2·6% (19 patients). Eighty-four ticks (11·3%) were PCR positive. Transmission occurred to 16 (26·7%) of 60 patients who were initially seronegative and could be followed up after the bite of an infected tick. These results indicate that the transmission rate from infected ticks in Europe is higher than previously assumed. Examination of ticks and antibiotic prophylaxis in the case of positivity appears to be indicated.

A multiethnic cohort of adult members of the Kaiser Permanente Medical Care Program (55300 men and 65271 women) was followed for 15 years (1979–93) to assess the association between total cholesterol and risk of infections (other than respiratory and HIV) diagnosed in the in-patient setting. Using multivariate Cox regression, total cholesterol was inversely and significantly related to urinary tract, venereal, musculo-skeletal, and all infections among men; and to urinary tract, all genito-urinary, septicaemia or bacteraemia, miscellaneous viral site unspecified, and all infections among women. The reduction of risk of all infections associated with a 1 s.d. increase in total cholesterol was 8% in both men (95% CI, 4–12%) and women (95% CI, 5–11%). For urinary tract infections among men, as for septicaemia or bacteraemia and nervous system infections among women, the risk relation was restricted to persons aged 55–89 years. Nervous system infections were positively related to total cholesterol among women aged 25–54. In both genders, the significant inverse association with all infections persisted after excluding the first 5 years of follow-up. Collectively, these data are suggestive of an inverse association, although not entirely consistent, between total cholesterol and incidence of infections either requiring hospitalization or acquired in the hospital. Further research is needed to elucidate whether these associations are biologically plausible or represent uncontrolled confounding by unmeasured risk factors.

One hundred and six Yersinia enterocolitica serogroup O[ratio ]3, biotype 4 isolated from human and porcine samples in 1984 and in the years 1993–5 were examined by pulsed-field gel electrophoresis (PFGE). The genomic profiles produced by the enzymes NotI and XbaI were studied. Sixteen (A–P) and 8 (1·8) different pulsotypes were obtained, respectively. By combining the pulsotypes produced by both NotI and XbaI 24 different types were distinguished. The two major types, designated as A1 and B1, comprised 36% of all strains tested. The proportions of pulsotypes A1 and B1 were, 35·9 and 25·6%, respectively, among strains isolated in 1984. The corresponding figures among the strains isolated in 1993–5 were 35·8 and 41·8%. Nine pulsotypes were found only in 1984 and nine only in 1993–5. The proportions of the major pulsotypes, A1 and B1, in human isolates were 42·9 and 35·7% and in porcine isolates 22·2 and 36·1% respectively. Six types were found among both human and porcine isolates, 8 only among human strains and 10 only among porcine strains.

Following the introduction of an improved surveillance system for infectious intestinal disease outbreaks in England and Wales, the Public Health Laboratory Service Communicable Disease Surveillance Centre received reports of 26 outbreaks between 1 January 1992 and 31 December 1995 in which there was evidence for waterborne transmission of infection. In these 26 outbreaks, 1756 laboratory confirmed cases were identified of whom 69 (4%) were admitted to hospital. In 19 outbreaks, illness was associated with the consumption of drinking water from public supplies (10 outbreaks) or private supplies (9 outbreaks). The largest outbreak consisted of 575 cases. In 4 of the remaining 7 outbreaks, illness was associated with exposure to swimming pool water. Cryptosporidium was identified as the probable causative organism in all 14 outbreaks associated with public water supplies and swimming pools. Campylobacter was responsible for most outbreaks associated with private water supplies. This review confirms a continuing risk of cryptosporidiosis from chlorinated water supplies in England and Wales, and reinforces governmental advice to water utilities that water treatment processes should be rigorously applied to ensure effective particle removal. High standards of surveillance are important for prompt recognition of outbreaks and institution of control measures. As microbiological evidence of water contamination may be absent or insufficient to implicate a particular water supply, a high standard of epidemiological investigation is recommended in all outbreaks of suspected waterborne disease.

To evaluate the seasonal trends of viral respiratory tract infections in a tropical environment, a retrospective survey of laboratory virus isolation, serology and immunofluorescence microscopy in two large general hospitals in Singapore between September 1990 and September 1994 was carried out. Respiratory tract viral outbreaks, particularly among infants who required hospitalization, were found to be associated mainly with respiratory syncytial (RSV) infections (72%), influenza (11%) and parainfluenza viruses (11%). Consistent seasonal variations in viral infections were observed only with RSV (March–August) and influenza A virus (peaks in June, December–January). The RSV trends were associated with higher environmental temperature, lower relative humidity and higher maximal day-to-day temperature variation. Although the influenza A outbreaks were not associated with meteorological factors, influenza B isolates were positively associated with rainfall. These data support the existence of seasonal trends of viral respiratory tract infections in the tropics.

A composite microporous membrane made of poly(N-benzyl-4-vinylpyridinium chloride) that showed strong affinity with bacterial cells was prepared as a filter material for removing airborne bacteria. Thickness, pore diameter and porosity of the membrane were 0·72 mm, 14·5 μm and 63%, respectively. Electron micrographic analysis revealed that the membrane consisted of a very large number of connected beads of 1·4 μm in diameter made of the pyridinium-type polymer. Filtration using the membrane was performed easily at low flow rates with insignificant pressure drop across the membrane. Filtration at 63·7 cm/sec gave 99·98% and 99·996% removal (3·7 and 4·4 log10-unit reduction in concentration) of Escherichia coli and Pseudomonas aeruginosa, respectively. Staphylococcus aureus was not detected in filtrates. Since pores of the membrane were much larger than these bacteria, the efficient removal was best explained in terms of the affinity of the polymer with bacterial cells.

The Escherichia coli flora of infants in developed countries is dominated by one or a few strains which persist for prolonged periods of time, but no longitudinal studies have been performed in developing countries. To this end, we studied the rectal enterobacterial flora in 22 home-delivered Pakistani infants during their first 6 months of life. Three colonies were isolated and species typed on each of 11 sampling occasions. E. coli isolates were strain typed using electromorphic typing of cytoplasmic enzymes, and their O serogroups were determined. There was a very rapid turnover of enterobacterial strains in the rectal flora of individual infants. On average, 8·5 different E. coli strains were found per infant, and several biotypes of other enterobacteria. Less than 50% of the infants were colonized with E. coli from their mothers, but strains of maternal origin were four times more likely to persists in the infants' flora than other E. coli strains. Enterobacteria other than E. coli were always of non-maternal origin, and Enterobacter cloacae and Klebsiella pneumoniae biotypes recovered from contaminated feeds were later identified in the infants' rectal flora. An early colonization with klebsiella or enterobacter was significantly associated with diarrhoea during the neonatal period, although these bacteria were not likely to be the cause of the disease. The results suggest that poor hygienic conditions result in an unstable and diverse enterobacterial flora, which may influence infant health.

We conducted a survey to determine the prevalence of known and theoretical exposure risks for cryptosporidiosis among selected New York City residents. Subjects were recruited from outpatients attending either a practice for persons with HIV infection (n=160), or other medical practices (n=153), at The New York Hospital–Cornell Medical Center. Despite a greater concern for waterborne infection, 82% of HIV-infected subjects reported consuming municipal tap water compared to 69% of subjects from other medical clinics (OR 2·1, 95% CI 1·2–3·6, P=0·006). Although 18% and 31% of subjects, respectively, denied any tap water consumption at home or work, all but one from each cohort responded positively to having at least one possible alternate source of tap water ingestion such as using tap water to brush teeth or drinking tap water offered in a restaurant. 78% and 76% of subjects, respectively, had at least one potential risk for exposure other than municipal water consumption, such as swimming in pools or contact with animals. Our findings indicate that it is possible to stratify the population into subsets by the amount of tap water consumed. This suggests that an observational epidemiologic study of the risk of contracting cryptosporidiosis from everyday tap water consumption is feasible.