Genotyping was used to analyse Pseudomonas aeruginosa isolates from sink drains and 15 intubated patients as part of a 3-month prospective study of strain transmission in a medical-surgical intensive care unit. Ninety percent of all washbasin drains were persistently contaminated with several P. aeruginosa genotypes. In 60% (9/15) of the patients, P. aeruginosa colonization or infection was hospital-acquired: P. aeruginosa strains isolated from these patients were present in hospital sinks or in other patients before their admission. Since all patients were immobile, personnel were the probable route of transmission of P. aeruginosa in the hospital. The mechanism of strain transmission from sinks to hands during hand washing was investigated in a children's hospital. When P. aeruginosa was present at densities of > 105/c.f.u. per ml in sink drains, hand washing resulted in hand contamination with P. aeruginosa via aerosol generation in the majority of experiments or P. aeruginosa was detected using an air sampler above the washing basin. High P. aeruginosa cfu were present at 4.30 h in the eight sinks (5.4 × 105−7.0 × 1010 c.f.u./ml), whereas at 13.00 h P. aeruginosa c.f.u. were significantly lower (3.1 × 102−8.0 × 105 c.f.u. / ml). These data reveal that the danger of bacterial contamination of hands during hand washing is highest in the morning. The identified transmission routes demand more effective hygienic measures in hospital settings particularly concerning personnel hands and sink drains.

One hundred and two Shigella spp. isolated in two hospitals in Hong Kong were analysed for antibiotic resistances, resistance plasmids and plasmid profiles. Three quarters of the isolates were S.flexneri. All isolates harboured plasmids, up to a maximum of ten within one strain. Plasmids of 220 kb encoding resistances to tetracycline, chloramphenicol and sulphonamide and probably also associated with invasivenes in the Sereny test were found in 80 strains and were transferable in 18% of cases. Resistance plasmids of 92 and 99 kb were found in 27 and 15 strains respectively and encoded resistances to ampicillin, tetracycline, chloramphenicol, kanamycin, sulphonamide, trimethoprim, cotrimoxazole and gentamicin; these plasmids were usually transferable. Four plasmids of 3·9, 2·8, 2·2 and 1·8 kb were commonly found in S. flexneri strains, but were rare in other species. In contrast, there was no predominant plasmid profile in S. sonnei. S. flexneri is endemic in Hong Kong and these plasmid studies suggest that the strains in circulation are derived from only a few clones.

The present study was undertaken to investigate the occurrence of antibiotic resistance in enteric flora in 64 children in rural Bangladesh over a 12-month period. The antibiotic resistance pattern of the isolates varied throughout the year and multiple resistance was highest during the post monsoon period. Seventythree percent of children had isolates resistant to more than three antibiotics throughout the year. Resistance to streptomycin was highest (78%), followed closely by ampicillin (72%). Of 82 multiply resistant isolates, plasmid DNA was demonstrated in 75%. Plasmid sizes ranged between 3·7 and 110 MDa, the commonest plasmids were of 70, 98 and 110 MDa. Complete or partial resistance was transferred by conjugation from 52% of the isolates, most frequently by single plasmids. The commonest plasmid incompatibility group was F11-A (46%) followed by incompatibility group P (22%). Plasmids of molecular weight 98 MDa most often hybridized with F11-A probes and those of 110 MDa with H11 probes. Plasmids from 10 transconjugants were digested with restriction enzymes and digest patterns demonstrated the presence of common plasmids. The findings show that there is a diverse, and mobile, genetic pool of resistance genes in this rural community. This genetic reservoir is potentially transferable to enteric pathogens, with major implications for public health and diarrhoeal disease control.

Plasmids belonging to the FIme incompatibility group were found in seven different serogroups of multiply antimicrobial-resistant Escherichia coli isolated from patients with urinary tract infection (UTI) and living in south-east London. Although widespread in Salmonella spp., FIme plasmids have only previously been described in E. coli in a strain of serogroup O15 K52 H1 responsible for an extensive and protracted outbreak of invasive community-acquired infection in south-east London in 1986. Our findings suggest either a wider background occurrence of FIme plasmids in E. coli associated with UTI than previously reported or alternatively, the dissemination and subsequent molecular diversification of the FIme plasmid associated with the epidemic strain of serogroup O15 K52 H1.

Between 1989 and 1991 a total of 1092 Vero cytotoxin-producing Escherichia coli O 157 isolated in the United Kingdom were phage typed in the Laboratory of Enteric Pathogens (LEP). Twenty-three phage types was identified, the most frequent being types 2 (36·1%), 49 (29·6%), 1 (10·3%) and 4 (8·9%). Although isolations of O 157 VTEC have increased each year from 1 in 1982 to 532 in 1991, the predominant phage types have remained unchanged although the proportion of strains belonging to types 2 and 49 has increased. O 157 VTEC from 17 outbreaks were phage typed during this period with phage type 49 predominating (7 of 17 outbreaks).

Over the 7 years 1985–91, 997 strains of Staphylococcus aureus from 962 patients with diseases other than food poisoning have been tested for the production of enterotoxins and toxic shock syndrome toxin-1 (TSST-1) and phage typed. In all, 128 cases could be classified as confirmed or probable toxic shock syndrome (TSS) but a further 199 cases were classified as possible or unconfirmed TSS. In 219 cases, an alternative diagnosis could be supported and 45 cases were classified as sudden infant death syndrome. In 371 cases, insufficient information for classification was available.

Strains of phage group I producing TSST-1 were associated with menstrual TSS. Many menstrual TSS cases were aged less than 20 and were using non-introducer tampons.

When all strains were reviewed, strong associations were observed between TSST-1 production and phage group I strains, enterotoxin B production and group V strains, enterotoxin C and phage-type 95 strains and between enterotoxin A without TSST-1 and phage group III strains.

Cholera vibrios can be allocated to one of three biotypes (classical, intermediate and El Tor), each of which can be sub-divided into two serotypes (Ogawa and Inaba). Vibriocidal tests with absorbed antisera have shown no evidence of biotype specificity in the killing of bacteria, but they have confirmed the role of the two serotype-specific antigens in immunity to cholera. The same presence of serotype specificity, and absence of biotype specificity, has been found by bacterial agglutination, an easier and quicker serological test. The use of this simpler test in ensuring a balanced serotype response to cholera vaccine is discussed, together with evidence that may lead to the production of more effective vaccine and better immunity.

Necrobacillosis occurs in man and animals. The typical forms of the disease in animals are caused by Fusobacterium necrophorum biovar A; biovar B strains are much less pathogenic. In this study the pathogenicity for mice of eight human isolates of F. necrophorum was compared with that of animal biovar A and B strains.

By subcutaneous inoculation seven of the human strains differed from biovar A but resembled biovar B in (1) producing, at the most, mild local lesions that rapidly healed, and (2) showing no enhancement of infectivity when suspended in sub-lethal doses of Staphylococcus aureus broth culture. The eighth human strain (A2433) resembled biovar A but differed from biovar B in (1) producing severe lesions, and (2) showing greatly enhanced infectivity in the presence of S. aureus. Nonetheless, strain A2433 differed from biovar A, both in the nature of the lesions produced and in its failure to cause severe general signs of illness and rapidly fatal infection.

By intravenous inoculation one of two biovar B strains and all except one of the eight human strains produced purulent lesions, often severe, in the liver and elsewhere, but infection was not usually associated with general signs of illness. In contrast, intravenous injection of a biovar A strain gave rise to a rapidly fatal infection, with severe lesions in the liver or elsewhere.

The results suggest that the term ‘necrobacillosis’ as used in human and veterinary medicine refers to diseases that differ in important respects.

Toxigenic strains of Staphylococcus aureus have been suggested to play a role in sudden infant death syndrome (SIDS). In this study we examined two factors that might enhance binding of toxigenic staphylococci to epithelial cells of infants in the age range in which cot deaths are prevalent: expression of the Lewisa antigen and infection with respiratory syncytial virus (RSV). By flow cytometry we demonstrated that binding of three toxigenic strains of S. aureus to cells from non-secretors was significantly greater than to cells of secretors. Pre-treatment of epithelial cells with monoclonal anti-Lewisa or anti-type-1 precursor significantly reduced bacterial binding (P < 0·01); however, attachment of the bacteria correlated only with the amount of Lewisa antigen detected on the cells (P < 0·01). HEp-2 cells infected with RSV bound significantly more bacteria than uninfected cells. These findings are discussed in context of factors previously associated with SIDS (mother's smoking, bottle feeding and the prone sleeping position) and a hypothesis proposed to explain some cases of SIDS.

Between 1969 and 1990 strains of Staphylococcus aureus from 359 outbreaks and sporadic cases of staphylococcal food poisoning in the United Kingdom were examined in the PHLS Food Hygiene Laboratory for the production of enterotoxin. In a number of instances the incriminated foods were also examined for the presence of enterotoxin. Strains from 79% of incidents produced enterotoxin A alone or together with another enterotoxin. The level of S. aureus present in the foods ranged from no viable S. aureus detected to 1.5 × 1010 c.f.u./g with a median of 3.0 × 107 c.f.u./g. Enterotoxin was detected in foods in the absence of viable S. aureus in only two outbreaks and in both cheese was the implicated food. Meat. poultry or their products were the vehicle in 75% of incidents with ham and chicken most frequently implicated. Other foods included fish and shellfish (7%) and milk and milk products (8%). Most contamination took place in the home followed by restaurants and shops. Seventy-one percent of the incident strains were lysed by phages of group III or I/III.

This study was undertaken to determine the distribution of phage types of Staphylococcus aureus isolates from hospital outbreaks or sporadic cases received in our laboratory during the past 14 years. The records for 15803 isolates from 55 Spanish hospitals have been analysed.

In relation to sporadic isolates we have been able to detect the predominance of phage group I and non-typable staphylococcal strains. Since 1989, we have observed a considerable increase in hospital infection caused by methicillin-resistant S. aureus strains which we could differentiate in to two groups; one belonging to phage group III (6/47/54/75/77/84/85) and other groups of non-typable strains which could be classified as phage group I-III after heat treatment (29/77/84) and with similar patterns by reverse typing (6/47/53/54/75/83A/84/85/W57/1030/18042).

During 1990 and 1991, these strains have extended widely to at least six different autonomous regions creating an epidemic situation in Spain.

In July 1989, 1834 samples of paté (of which 1698 were from retail displays) were examined by the PHLS for the presence of Listeria monocytogenes. The survey was repeated in July 1990, when 626 paté samples on retail sale were examined. Between the two surveys there was a marked reduction in the proportions of patés contaminated (10% in 1989 and 4% in 1990) and in the numbers of samples from which > 103L. monocytogenes/g were recovered. The higher rate of contamination detected in 1989 was largely due to paté from a single manufacturer. In both surveys, paté sold as loose slices had higher rates of contamination than those prepackaged. Temperature control had improved between the two surveys where 65% of samples in 1989 and 83% in 1990 were stored at ≤ 7 °C. Although contamination occurred at almost all temperatures, L. monocytogenes was both quantitatively and qualitatively more common in samples stored at > 7 °C. The majority of patés had unexpired shelf lives of between 0 and 3 weeks. Although contamination occurred throughout the shelf life of these products, the proportion of samples where L. monocytogenes was recovered was higher in patés with expired sell by dates. There was an association between high total viable counts and the presence of L. monocytogenes.

Salmonella typhimurium is the most frequently identified serovar of Salmonella in Italy. This serovar is characterized by the widespread dissemination among human and non-human sources of phenotypically and genetically well-differentiated clones.

In this study 457 strains of S. typhimurium isolated in Italy in the years 1982–91 from human and animal sources were submitted to characterization by the rDNA fingerprinting technique. Application of this typing method, after digestion of chromosomal DNA with HincII endonuclease, confirmed the greatest genetic differentiation of clones of S. typhimurium, allowing reliable identification of 45 rDNA patterns linked into 9 major clusters. rDNA pattern clusters or ribotypes specific to man were not recognized, whereas some rDNA patterns were characteristically related to ducks, pigeons and pet birds. The ribotyping results for isolates from animal hosts suggest that pig and cattle are the main source of human infection.

Routine surveillance of infection in England and Wales detected 394 cases of campybacter bactermia in 11 years. This represented an average incidence of 1.5 per 1000 intestinal campllobacter infections, with range of 0.3/1000 in children aged 1–4 years to 5.9/1000 in patients aged 65 years or more. Definitive identification of 257 isolates showed that 89% were Campylobacter jejuni or C. coli: other species were C. fetus (8.6%). C. lari (0.8%), C. uspaliensis (0.8%). helicobacter (Campylobacter) fennlliae (0.8%), and Helicobacter (Camphlobacter) cinaedi (0.4%). Most (71%) of the C. jejuni/C. coli bacteraemias were in pastients with acute enteritis. Of the patients with C. fetus bacteremias only 27% had diarrhoea: they were older than patients with diarrhoae: they were older than patients with C. jejuni or C. coli bacteremia (54.1 r. 45.9 years) and proportionally more of them were male (M:F retio 2.7: 1 v. 1.9:1): 41% had endovascular pathology or cellulitis. There was a higher proportion of C. jejuni serogroup O 4 (Penner) and O 18 strains among blood than faecal isolates. Which suggests that they were unsually serum resistant and/or invasive.

A total of 79 Australian isolates of beta-haemolytic Escherichia coli from cases of porcine postweaning diarrhoea (PWD), and 18 isolates of serotype O 149:K91:K88 (F4) from unweaned pigs from Australia, Indonesia and Denmark, were examined by multilocus enzyme electrophoresis. These were divided into 57 electrophoretic types (ETs), with an overall mean genetic diversity per enzyme locus of 0·466. This value closely resembled that previously recorded for the whole species. Not only was the collection diverse, but there was considerable genetic heterogeneity amongst PWD isolates of the same serogroup. Isolates from serogroups O 8 and O 138 were most varied, whilst many from serogroups O 141 and O 149 were more closely related. In contrast, the isolates from the unweaned pigs all belonged to only one ET.

An outbreak of non-bacterial food poisoning presumed due to small round, structured viruses (SRSV) occurred at a national conference.

A detailed postal survey of all conference attenders was carried out to ascertain the cause of the outbreak and 355 questionnaires were returned.

Univariate analysis showed that mussels in the seafood cocktail were the likely vehicle of infection. A dose–response relationship between the amount of seafood cocktail consumed and the risk of illness was demonstrated. Dose–response has not previously been documented in a food-borne outbreak due to small round structured virus.

Detailed quantitative food histories can be useful in eliciting dose-response relationships and may be crucial in establishing the vehicle of infection when investigating food poisoning following consumption of a set-menu meal. Their use should be considered in other outbreak situations.

This survey reports the results of investigations performed by the Laboratory of Enteric Pathogens (LEP), to identify evidence of human infection with Vero cytotoxin-producing Escherichia coli (VTEC) in the UK during the period 1989–91. Bacterial isolates, faecal specimens and serum samples were received from patients suffering from diarrhoea, bloody diarrhoea and haemolytic uraemic syndrome. Using serotyping, Vero cytotoxin gene probing and an ELISA for serum antibodies to E. coli 0 157, evidence of infection was detected in 232, 428 and 615 individuals in 1989. 1990 and 1991 respectively. Of these individuals, 15% were reported as having HUS. Vero cytotoxin-producing E. coli O 157 was the most frequently encountered serogroup. with isolations from a total of 1092 individuals over the 3-year period. The incidence of VTEC infection increased from 0·41/ 100000 in 1989 to 1· 07/100000 in 1991. The area with the highest rate of infection in each year was Scotland, increasing from 1–37/100000 in 1989 to 3·97/ 100000 in 1991.

Between June 1990 and July 1991, broiler chickens from 49 flocks from 23 farms were examined for the carriage of Campylobacter jejuni at slaughter. Thirty-seven flocks (76%) were campylobacter-positive. Prevalence of campylobacter-colon-ization was not associated with any of a variety of factors such as water source and broiler house floor structure. There was also no apparent seasonal variation in carriage. Investigations on one farm indicated that dipping boots in disinfectant before workers entered broiler houses either delayed or prevented colonization with C. jejuni.

The occurrence and nature of bacteriologically confirmed tuberculosis due to Mycobacterium tuberculosis in South-East England in the period 1984–91 is reported and compared with the results of a study for 1977–83. Registered new cases reached a low of 1028 in 1988 but increased to 1252 in 1991. This appeared to be due to a halt in the previous decline in new cases of European patients, a small increase in the number of Indian subcontinent (ISC) patients and an increase in patients from Africa. A total of 122 patients, mostly of European ethnic origin, were known to be HIV positive. As in the 1977–83 study, disease in the ISC group affected younger patients than in the European group, tubercle bacilli were more frequently isolated from a non-pulmonary site in the ISC group (45%) than in the European group (19%) and there was a higher incidence of the South Indian variant of M. tuberculosis in the former group (17%) than in the latter (9%). The overall incidence of drug resistance has not altered significantly since the 1977–83 study but 46 strains resistant to 3 or more drugs were isolated from 4099 ISC patients, compared with 3 of 4594 strains from European patients. Six of the 122 isolates from HIV positive patients were drug-resistant but none was multi-drug resistant.

The slight rise in registered bacteriologically proven cases of tuberculosis, the presence of multi-drug resistant strains, the occurrence of HIV-related tuberculosis and reports of the emergence of multi-drug-resistant HIV-related tuberculosis in other countries strongly indicate the need for continued careful surveillance.

The presence of pneumococcal capsular antigen (PCA) in the oropharynx was sought in subjects without respiratory tract infection. Saliva specimens from 239 subjects were analysed by counter-current immunoelectrophoresis using ‘Omni-serum’. 15.5% gave positive reactions but only 24% of positive samples were typable and therefore due to pneumococcal or pneumococcal-like antigens. Given that oropharyngeal production of antigens occurs we investigated whether PCA in expectorated sputum arose from oropharyngeal contamination. Sixteen patients with pneumococcal pneumonia, and with sputum positive for PCA, were investigated in detail. On the basis of serotyping and concentration the PCA in sputum was thought to arise from the lower respiratory tract in all cases. This was confirmed by a simple, novel approach involving the comparison of concentrations in concomitant samples of saliva and sputum. Thus while oropharyngeal production of antigens poses a potential diagnostic problem the latter approach can be used to exclude contamination.