This study evaluated the epidemiology of Pseudomonas aeruginosa keratitis in contact lens (CL) wearers; the relationships between CL storage case contamination and CL hygiene practice and between CL hygiene and the development of keratitis. Sixteen CL wearers with keratitis were compared with 44 asymptomatic controls. Lens hygiene practice was assessed and CL care materials, domestic water sites and endogenous sites were evaluated microbiologically. Poor CL hygiene was not associated with Ps. aeruginosa keratitis. There was an association between keratitis and bacterial contamination of the CL and storage case (P < 0·0005). Lens and storage case contamination were not significantly associated with poor hygiene. No domestic or endogenous source for Ps. aeruginosa was found. Causative organisms may be derived from other sources, but CLs and CL storage cases provide a favourable environment for Ps. aeruginosa colonization. Changing the CL care environment to one less favourable for Ps. aeruginosa may help to eliminate this problem.

Using pulsed-field gel electrophoresis (PFGE) and 16S rRNA (rrn) analysis (ribotyping). the in viro derivation of strains of Salmonella enteritidis PTs 9a and 7 from a strain of S. enteritidis PT 4 has been demonstrated. All strains were isolated from a single patient over a 6-week period. Further studies have demonstrated that in terms of pulsed-field profile and ribotype, the genotypes of the patient-derived strains differed from those of the reference strains of the respective phage types. It is concluded that when used in combination, these methods can provide evidence of phylogenetic relationships in apparently unrelated S. enteritidis phage types isolated during pathogenesis of disease.

Broiler flocks on two Dutch poultry farms were screened weekly for the presence of Campylobacter in fresh caecal droppings during eight consecutive production cycles. Hatchery and fresh litter samples were taken at the start of each new cycle. Water, feed, insects, and faeces of domestic animals, present on the farms were also included in the sampling. Penner seretyping of isolates was used to identify epidemiological factors that contribute to Campylobacter colonization in the broiler flocks. Generally, broiler flocks became colonized with Campylobacter at about 3–4 weeks of age with isolation percentages of 100%, and stayed colonized up to slaughter. A similar pattern of serotypes was found within the various broiler houses on one farm during one production cycle. New flocks generally showed also a new pattern of serotypes. Most serotypes isolated from the laying hens, pigs, sheep and cattle were different from those isolated from the broilers at the same time. Campylobacter serotypes from darkling beetles inside the broiler houses were identical to the ones isolated from the broilers. No Campylobacter was isolated from any of the hatchery, water, feed or fresh litter samples. Conclusive evidence of transmission routes was not found, but results certainly point towards horizontal transmission from the environment. Horizontal transmission from one broiler flock to the next one via a persistent contamination within the broiler house, as well as vertical transmission from breeder flocks via the hatchery to progeny, did not seem to be very likely.

Several typing systems have been described for Campylobacter jejuni and C. coli, to assess the complex epidemiology of these important enteric pathogens. In the present study two typing methods, slide agglutination according to the Lior scheme, and the demonstration of restriction-fragment length polymorphisms (RFLP) of flagellar genes, have been used in parallel on a set of 194 strains. This set comprised 118 sero-reference strains of C. jejuni and C. coli of the Lior scheme, as well as 76 clinical isolates. All isolates were serotyped and subjected to PCR for amplification of flagellar genes, and the PCR product was restricted with Alu I. Flagellar genes could be amplified in 152 strains. Among 85 seroreference strains, 74 different RFLP patterns were observed, and among 67 clinical isolates, there were 36 patterns. There was only limited correlation between flagellar RFLP and the Lior serogroup, and the variability of patterns in serogroups HL2 and HL4 were as marked as the variability between serogroups. Flagellar gene RFLP patterns are shown to be stable, highly discriminatory epidemiologic markers.

Thirty-six encoded isolates of Escherichia coli, 32 of which were of serotype O157. were examined by pyrolysis mass spectrometry (PyMS). Thirty-one of the serotype O157 isolates possessed the flagellar antigen H7 and produced Verocytotoxin (VT). the other isolate serotyped as H45 and was non-toxigenic. Eighteen of the VT-producing E. coli (VTEC) isolates were from sporadic disease in residents of the Northern Region. Standard principal component (PC) and canonical variate (CV) analysis of the data distinguished only the four non-O157 isolates from the remainder which were indistinguishable by this approach. A similarity matrix based on differences between individual CV means distinguished a further ten isolates. The matrix correctly clustered 2 pairs of isolates from siblings and 4 isolates from an affected family. A further 5 clusters of 3 or more isolates and 6 pairs of isolates were defined. These groupings proved to be homogenous for toxin phenotype but occasionally entrained isolates of dissimilar phage type. However, in general, PyMS-derived clustering of apparently sporadic isolates accorded with geographical locations as determined by postcode. PyMS, which is a quick and high volume capacity phenotypic technique. may be a useful addition to existing methods in the investigation of the epidemiology of sporadic VTEC disease.

A summer outbreak of severe gastroenteritis followed by haemolytic uraemic syndrome (HUS) and thrombotic thrombocytopenic purpura in a nursery school and kindergarten is described. Sandwiches prepared with green butter made with contaminated parsley were the likely vehicle of infection. The parsley originated from an organic garden in which manure of pig origin was used instead of artificial fertilizers, Cornally identical Verotoxinogenic Citrobacter freundii were found as causative agents of HUS and gastroenteritis and were also detected on the parsley.

A 2-year case-control study was conducted to describe the epidemiology of Haemophilus influenzae type b (Hib) and investigate Hib disease risk factors in the Republic of Ireland. Between October 1991 and September 1993, 149 laboratory confirmed incident cases were matched with community controls. Annual Hib disease incidence was 25·4 per 100000 children under 5 years, with peak incidence (65·8 per 100000) in the 6–11 months age-group. Meningitis was the predominant clinical condition. Twenty-four (16·1%) isolates were resistant to ampicillin. Crèche or day-care attendance and the presence of chronic illness emerged as risk factors for Hib disease. Empirical first line treatment for suspected Hib infection warrants alternatives to ampicillin such as cefotaxime. Completed immunization with Hib conjugate vaccine by 6 months of age is required for maximum disease prevention. Until all children are receiving Hib vaccine on schedule, those who are crèche or day-care attendees and those with chronic illness should be prioritized for timely immunization.

The prevalence of antibodies to Chlamydia pneumoniae was examined in a stratified random sample of 581 Jerusalem adult residents between August 1987 and March 1989. IgG and IgA titres were measured by microimmunofluorescence, and associations with smoking and socio-demographic variables were assessed. IgG antibodies were found in 84·5% (95% confidence interval (CI): 80·4–87·9) of men and 68·7% (95% CI: 61·6–75·0) of women (P < 0·0001 for sex difference), indicating a very high rate of exposure in this population. IgA antibodies, postulated to represent persistent infection, were present in 45·1% (95% CI: 40·1–50·2) of men and 23% (95% CI: 17·4–29·7) of women (P < 0.0001 for sex difference). Factors associated with IgG seropositivity included family size, education and social class. On the other hand, age (in men) and smoking were associated with IgA seropositivity. These findings support the hypothesis that low socioeconomic status and household crowding may be predictive of exposure to or infection with this organism (IgG seropositivity), whereas they do not explain persistence of the infection putatively expressed as IgA seropositivity.

To investigate the feasibility of using a record linkage method for identifying vaccine attributable adverse events, computerized hospital admissions and vaccination records from South East Kent district were linked and checked for accuracy. Records for 90% of children under 2 years of age admitted to hospital over a 2-year period were matched with vaccination records using a computer algorithm based on name, date of birth, sex, and post-code supplemented by visual inspection. Relative to this gold standard, matching on date of birth, sex and postcode alone had a sensitivity of 60% and an incorrect match rate of 0·2% after matches to more than one vaccine recipient were excluded. Manual checking of a sample of admissions showed that only 4% had been assigned incorrect International Classification of Disease (ICD) codes. Routine record linkage of ICD admission codes to vaccination records therefore yields data of good quality which may be used for surveillance purposes.

To examine the extent of wild poliovirus circulation during the 1992–3 epidemic in the Netherlands caused by poliovirus type 3, 269 samples from sewage pipelines at 120 locations were examined for the presence of poliovirus. The epidemic virus strain was found in 23 samples, all from locations inside the risk area which contained communities that refuse vaccination for religious reasons. By sewage investigation, the wildtype virus was shown to be present in the early phase of the epidemic at two locations, one week before patients were reported from that area. The wild type 3 poliovirus was also detected retrospectively in a river water sample collected for other reasons three weeks before notification of the first poliomyelitis case, at a site a few kilometres upstream the home village of this patient. Oral poliovirus vaccine (OPV) virus was found at 28 locations inside or at the border of the risk area. Trivalent OPV was offered to unvaccinated or incompletely-vaccinated persons living in this region as part of the measures to control the epidemic.

The sequence variation in a 934 base-pair region of the gene encoding the haemagglutinin-neuraminidase of five human parainfluenza virus type 3 (HPIV3) isolates was determined together with that of a prototype UK strain. All of the clinical isolates were from the Manchester area of the UK and were obtained in 1990. 1991 and 1993. The gene segment was amplified by the polymerase chain reaction using HPIVB-specific oligonucleotide primers. The nucleotide homology of the strains was high, around 99% and specific differences in the UK sequences when compared with that of the US prototype strain were identified. In addition, a number of isolate-specific differences were seen. No correlation was detected between the observed nucleotide mutations and the year of isolation, which supports the hypothesis that HPIV3 shows cocirculation of a heterogeneous population of viruses rather than varying with time in a linear fashion. However, the data suggested that geographically-defined genetic lineages of HPIV3 may exist.

Of eight acute infections in German tourists caused by sandfly fever virus, serotype Toscana (TOS). and diagnosed clinically and serologically, seven were acquired during visits to Tuscany, Italy, and one to Coimbra, Portugal. An indirect immunofluorescence assay (IFA) using infected cells, and a newly developed enzyme-immunoassay (EIA) using crude virus antigen prepared from infected Vero-E6 cells was used to detect anti-TOS IgM and IgG. In a seroepidemiological survey of 859 health care workers and medical students, anti-TOS IgG was detected in 1·0% by IFA, and in 0·7% by EIA. In 2034 German patients hospitalized for various diseases, 1·6% were positive for anti-TOS IgG by IFA, and 0·8% by EIA. Anti-TOS IgG was detected in 43 samples of commercial immunoglobulins at titres of 10–1000 by EIA. Although the seroprevalence of antibodies to TOS is low in Germany, TOS infection should be considered in patients returning from endemic areas who complain of fever, and headaches, and have symptoms of meningitis.

Samples from a sow serum bank representative of the pig population of Great Britain collected during 1991–2, were examined for antibodies to influenza A, B and C viruses, using viruses which had been isolated from a variety of hosts. For influenza A viruses there was evidence of the continued circulation of ‘classical swine’ H1N1 virus (26%) seroprevalence), and human H3N2 viruses (39%) which are antigenically most closely-related to A/Port Chalmers/1/73 virus. In addition antibodies were detected to A/swine/England/201635/92 (8%), a strain of H3N2 virus which appears to have arisen by antigenic drift from conventional H3N2 swine strains. Specific antibodies (2%) were detected to an H1N1 virus (A/swine/England/195852/92) related most closely to avian H1N1 strains. In tests with human H1N1 and H3N2 viruses, excluding isolates from pigs, the highest seroprevalence was detected to the prevailing strains from the human population. Serological tests with avian H4 and H10, human H2, equine 1 and 2 influenza A viruses were all negative. Seven pigs seropositive by haemagglutination-inhibition, virus neutralization and immunoblotting assays for antibody to influenza B virus, were randomly distributed geographically suggesting that influenza B viruses may be transmitted to pigs but fail to spread. The seroprevalence to influenza C viruses was 9·9% indicating that these viruses are widespread in pigs. These results provide further evidence that the pig can be infected by a number of influenza viruses, some of which may have significance in the epidemiology of human influenza.

Epidemiol. Infect. 113 (1994), 501–11

J. E. Childs, C. V. Trimarchi and J. W. Krebs

The epidemiology of bat rabies in New York State, 1988–92

The words ‘or animal’ in the summary preceding the word ‘contact’ (page 501) and the words ‘or domestic pets’ from the first sentence in the section ‘Reported bat bites, contacts and rabies’ on page 504 should be omitted. The data base included only human contacts with bats and any reference to domestic animal contact is incorrect. The authors regret this oversight.