Traditional enteropathogenic Escherichia coli of infantile diarrhea

First described in 1885, Escherichia coli gradually emerged as a cause of infantile diarrhea. Currently, four pathogenic categories of diarrheagenic E. coli are recognized: enteropathogenic (EPEC), enterotoxigenic (ETEC), enteroinvasive (EIEC), and enterohemorrhagic (EHEC). Of these, ETEC and EPEC are important causes of pediatric diarrhea, especially in developing countries. ETEC strains are characterized by their production of well-defined heat-labile or heat-stable enterotoxins. In the absence of other identifying characteristics, EPEC strains have been identified by serotyping only; recent laboratory and clinical investigations have shown that these strains constitute a distinctive group of pathogenic microorganisms displaying characteristic adherence to cultured epithelial cells and producing pathognomonic histopathologic changes in experimentally infected animals and naturally infected children. The pathogenicity of most strains of EPEC is associated with the carriage of a 55- to 70-megadalton plasmid that specifies the production of an adherence factor. EPEC strains may also owe their pathogenicity to the production of at least one enterotoxin, possibly a Shiga-like cytotoxin.

Production of aerobactin by some species of the genus Yersinia

Seventy-seven isolates of Yersinia species were examined for evidence of hydroxamate biosynthesis. A total of 7 of 12 Y. frederiksenii, 1 of 5 Y. kristensenii, and 5 of 5 Y. intermedia isolates synthesized an hydroxamate species which was chromatographically and electrophoretically identical to aerobactin. Proof that this material was indeed aerobactin was obtained for one strain of Y. frederiksenii by using mass spectrometry. None of 50 Y. enterocolitica nor any of 5 Y. pseudotuberculosis isolates produced hydroxamates.

Influence of the virulence plasmid and the Congo red reaction on the antimicrobial susceptibility of Yersinia species

Seventeen strains of Yersinia enterocolitica and related species were examined for their susceptibility to a variety of antimicrobial agents. Twelve of the strains were clinical isolates which carried a virulence plasmid; five were food isolates which lacked plasmids. When grown on agar containing Congo red or haemin, each strain yielded pigmented (CR+) and non-pigmented (CR-) variants. CR+ bacteria of clinical origin were virulent, whereas CR+ food isolates and all CR- bacteria were avirulent. The susceptibility of CR+ and CR- bacteria to the following agents was compared: ampicillin, chloramphenicol, latamoxef, nalidixic acid, novobiocin, polymyxin B, rifampicin, streptomycin, tetracycline and tobramycin. The results showed that bacteria which bound Congo red were more susceptible to each drug examined than their CR- counterparts. This relationship held regardless of whether the CR+ bacteria possessed a virulence plasmid or not. The findings suggest that binding of antimicrobial agents by yersiniae may parallel binding of haemin and Congo red.

In vitro assessment of virulence in Yersinia enterocolitica and related species

We have examined 136 isolates of Yersinia species, comprising 112 strains of Yersinia enterocolitica, 12 of Y. frederiksenii, 8 of Y. intermedia, and 5 of Y. kristensenii, for the presence of 40- to 50-megadalton virulence-associated plasmids and expression of the following plasmid-associated characteristics: Congo red pigmentation (CR), calcium dependence, autoagglutination, hydrophobicity, resistance to normal human serum, and pathogenicity in mice. All 136 strains yielded both pigmented (CR+) and nonpigmented (CR-) variants. Only CR+ variants, however, were virulent for iron-overloaded, desferrioxamine B-treated mice (R. M. Robins-Browne and J. K. Prpic, Infect. Immun. 47:744-779, 1985). Although the in vitro virulence-associated characteristics generally occurred together, each one could be expressed independently. Strains of Y. frederiksenii, Y. intermedia, and Y. kristensenii also expressed individual virulence-associated properties. Of 53 Y. enterocolitica strains which were virulent for iron-overloaded, desferrioxamine-treated mice, all but one expressed every virulence-associated characteristic. Several strains which were avirulent for mice, however, demonstrated these characteristics in various combinations. Because many Yersinia strains, particularly environmental isolates, carried plasmids of 40 to 50 megadaltons, detection of plasmids provided little information about bacterial pathogenicity unless virulence-associated properties were also sought. The best in vitro predictor of virulence was autoagglutination, followed by calcium dependence. Because only CR+ variants expressed virulence-associated determinants, Congo red pigmentation is useful for selecting potentially virulent strains.

The pathogenesis of Yersinia enterocolitica infection in gnotobiotic piglets

Yersinia enterocolitica is an important cause of enteritis and mesenteric adenitis in many countries. However the pathogenesis of the disease caused by this organism has not been fully elucidated. Most isolates from clinical material possess two independent properties associated with virulence whose relative contribution to the development of disease is not known. These are the ability to penetrate the intestinal wall, which is thought to be controlled by a plasmid gene, and the production of heat-stable enterotoxin, which is controlled by a chromosomal gene. In this study, we infected neonatal gnotobiotic piglets with strains of Y. enterocolitica expressing these two properties in various combinations. The suitability of the piglet model was shown in experiments in which piglets fed virulent Y. enterocolitica serogroup O3 developed a clinical illness related to the size of the inoculum, which was accompanied by intestinal lesions similar to those reported in naturally and experimentally infected people and animals. The results confirmed the key role of a 47 X 10(6)-mol. wt plasmid in the pathogenicity of Y. enterocolitica, but suggested that penetration of the intestinal wall may be governed by chromosomal rather than plasmid-borne genes. No role for enterotoxin in the pathogenesis of yersiniosis was shown, although there was evidence that enterotoxin may promote intra-intestinal proliferation of Y. enterocolitica, thus favouring increased shedding of bacteria and encouraging their spread between hosts.

Enteropathogenic Escherichia coli enteritis: evaluation of the gnotobiotic piglet as a model of human infection

The pathogenicity of classical enteropathogenic Escherichia coli strains of human origin was investigated in gnotobiotic piglets. One to two day old piglets in groups of four were infected perorally with approximately 10(8) colony forming units of one of eight enteropathogenic E coli strains or a non-pathogenic control strain. Animals were necropsied 24 or 48 hours after infection and their intestines were subjected to histological examination, quantitative bacterial culture and estimation of lactase activity. Four enteropathogenic E coli strains caused mild to moderate diarrhoea in nine of the 16 piglets inoculated with them. Piglets given two of these strains later became moribund. One enteropathogenic E coli strain caused a severe illness unaccompanied by diarrhoea. Inflammation of the intestinal mucosa occurred with all eight enteropathogenic E coli strains, but not with the control strain. Pathological changes were most pronounced in the distal ileum and colon and adherent bacteria were seen on the surface of the inflamed mucosa. The extent of the inflammatory response in infected piglets for the most part paralleled the severity of the clinical signs, the degree of bacterial colonisation and the reduction in lactase activity. Electron microscopic examination of tissue from piglets infected with three different strains showed that bacterial adherence to the apical plasma membrane of epithelial cells was accompanied by distinctive ultrastructural changes. These included degeneration of the microvillous brush border, together with cupping and pedestal formation of the plasma membrane at sites of bacterial attachment. The same changes have been seen in naturally occurring enteropathogenic E coli diarrhoea in humans and rabbits. The combined clinical and pathological findings indicate that the neonatal gnotobiotic piglet is a suitable model of infection with enteropathogenic E coli.

Effects of iron and desferrioxamine on infections with Yersinia enterocolitica

The effects of iron-dextran and the iron chelator desferrioxamine B mesylate (Desferal) on the course and outcome of experimental yersiniosis were investigated. Yersinia enterocolitica strains representing the three leading serogroups pathogenic for humans, O3, O8 and O9, were studied. In mice, iron-dextran reduced the median lethal dose of intraperitoneally administered Y. enterocolitica O3 and O9 ca. 10-fold, whereas Desferal reduced this value more than 100,000-fold. Experiments in which Y. enterocolitica was given orally to mice and intraconjunctivally to guinea pigs confirmed that Desferal markedly increased the susceptibility of animals to yersiniosis. Although serogroup O8 yersiniae were inherently more virulent for laboratory animals, they were less affected by Desferal than were O3 or O9 strains. In vitro experiments indicated that Desferal promoted growth of Y. enterocolitica under iron-limiting conditions and suggested that the enhanced virulence of O8 yersiniae may be due to their comparatively low requirement for iron. The adverse effect of Desferal on the course of experimental infection with Y. enterocolitica may partly explain the heightened susceptibility of iron-overloaded patients to systemic yersiniosis.

Microbiological studies during the perinatal period. An attempt to correlate selected bacterial and viral infections with intra-uterine deaths and preterm labour

An investigation of the genital, amniotic fluid, placental and neonatal microflora in a population with a high prevalence of amniotic fluid infections and perinatal deaths is reported. Women presenting with preterm labour and intra-uterine death were investigated for the presence in vaginal secretions of a selected variety of pathogenic organisms, and the findings in these patients were compared with those in three groups of controls who presented with normal labour at term or late delivery or who required elective caesarean section as a result of cephalopelvic disproportion. Indirect evidence of infection was also sought by assaying cord blood for immunoglobulin levels and by testing for specific antibodies to well-known congenitally acquired pathogens. The results showed a significant relationship between gonococcal infection, serological evidence of Chlamydia trachomatis infection and preterm labour.

Antibiotic-resistant pneumococci in hospitalized children

A search for nasopharyngeal carriers of Streptococcus pneumoniae was conducted in 573 children hospitalized in Durban, South Africa. Study subjects were divided into two groups, comprising 305 new admissions and 268 patients who had been hospitalized for more than 24 h. Of the 573 children 178 (31%) yielded pneumococci on nasopharyngeal culture; 99 (32%) and 79 (29%) children in the new admission and in-patient categories respectively. Twenty-one (12%) pneumococci were resistant to penicillin, including 11 strains that were resistant to more than one antibiotic. Resistant pneumococci belonged exclusively to serotypes 6 and 19 (Danish nomenclature), which were also the commonest serotypes among penicillin-sensitive strains. Factors that correlated with carriage of penicillin-resistant pneumococci were hospitalization for more than 24 h, young age and recent exposure to beta-lactam antibiotics.

Aetiology of diarrhoea in adequately nourished young African children in Durban, South Africa

Two inter-related hospital-based studies on the causes of diarrhoea in African children who were not overtly malnourished, were performed over a period of two years in Durban, South Africa. The first study involved 126 inpatients selected for previously untreated diarrhoea of less than 96 h duration. On examination and culture of the stools of these children bacteria were identified in 60%, rotavirus in 20% and parasites in 4%. No pathogens were identified in 33%. The bacteria most frequently encountered were Campylobacter jejuni (21%), enteropathogenic Escherichia coli (EPEC) (18%), enterotoxigenic Escherichia coli (ETEC) (11%) and salmonella (8%). Of those cases having an identifiable cause 16% had more than one organism. The majority of children had mild disease and recovered clinically in 5.4 (1.3) days (Mean (s.d.]. The second study included 352 outpatients with diarrhoea who were randomly selected without regard to duration of disease or prior treatment. The results of faecal investigations were as follows: 28% had bacterial pathogens, 34% rotavirus, 15% parasites and 61% no pathogens. The bacteria most frequently detected were C. jejuni 7%, EPEC 7%, salmonella 6%, and shigella 5%. Ten per cent of positive cases had more than one pathogen. The most likely reason for the higher percentage with no identifiable pathogen is the prolonged duration of diarrhoea in 19% for more than seven days. In 128 control children with diarrhoea, stool samples were investigated during the same period as the two studies: 13% had bacteria, 2% rotavirus and 14% parasites. The bacteria involved were C. jejuni (5%). EPEC (5%), salmonella (2%) and shigella (2%).

Seasonal and racial incidence of infantile gastroenteritis in South Africa

The etiology of summer diarrhea, which formerly caused extensive mortality in children living in industrialized countries, was never discovered. This condition no longer occurs in developed countries, having been replaced by winter diarrhea, which is associated with a low mortality. Summer epidemics of diarrhea still take place in black South African infants, whereas, in white South African children, the pattern of diarrhea is similar to that seen in children in industrialized countries today. In 1977-1980, the author studied the records of patients less than two years of age admitted for treatment of dehydrating diarrhea to two teaching hospitals in Johannesburg, the Johannesburg General and Baragwanath Hospitals, which serve the needs of white and black patients, respectively. The incidence of severe diarrhea (i.e., diarrhea necessitating hospitalization) showed highly significant seasonality (p less than 0.001) and distinctive patterns by race. Dehydrating diarrhea in black children was strongly associated with warm weather, while diarrhea in white children occurred more regularly throughout the year, with a peak incidence in late fall. Laboratory studies have shown that bacteria, in particular "classical" enteropathogenic Escherichia coli, are the leading cause of diarrhea in black South African children, and that diarrhea in white children is largely attributable to rotaviruses. The association of enteropathogenic E. coli with diarrhea in black children suggests that these bacteria were responsible for earlier outbreaks of summer diarrhea. The finding that the etiology of diarrhea varies according to socioeconomic class has important implications for diarrhea control programs.

Differentiation between virulent and avirulent Yersinia enterocolitica isolates by using Congo red agar

Cultivation of clinical isolates of Yersinia enterocolitica of diverse geographical origin on a medium containing 5 micrograms of Congo red per ml disclosed two colony types. These were designated CR+ and CR- according to their ability to bind Congo red. CR+ strains bore plasmids of between 40 and 50 megadaltons and were positive in several tests of Y. enterocolitica virulence, including autoagglutination, reduced growth on magnesium oxalate agar, resistance to the bactericidal effect of serum, and lethality for iron-overloaded mice. CR- strains were plasmidless and were negative in all these assays. The Congo red reaction provides a simple and efficient means of screening Y. enterocolitica for virulence and is the best available method for identifying individual plasmid-bearing colonies.

Treatment of acute nonspecific gastroenteritis of infants and young children with erythromycin

A double-blind placebo-controlled trial of erythromycin ethylsuccinate was conducted in 65 infants and young children hospitalized with acute nonspecific gastroenteritis. Etiologic agents included rotaviruses (29%), Campylobacter jejuni (17%), "classical" enteropathogenic Escherichia coli (12%), enterotoxigenic E. coli (11%), Salmonella (9%), Shigella (2%), and Giardia lamblia (2%). No pathogens were obtained from 25 (38%) children. Treatment with erythromycin had no effect on the course of the illness in terms of the time required for hydration, stool frequency and temperature to return to normal, or for vomiting to be abolished. Children treated with erythromycin, however, experienced a marginally, but significantly (P less than 0.05), shorter period of abnormal stool consistency compared with control subjects. This effect was most pronounced in children from whom no enteropathogens were isolated.

Treatment of Campylobacter-associated enteritis with erythromycin

Twenty-six infants and young children with acute dehydrating diarrhea associated with Campylobacter jejuni participated in a randomized, double-blind, placebo-controlled therapeutic trial. Of 25 patients who completed the study, 11 were treated for five days with oral erythromycin ethylsuccinate (40 mg/kg/day in divided doses), and the rest received matched placebo. Although erythromycin significantly shortened the duration of C jejuni excretion, it appeared to exert no effect on the clinical course of the illness. This failure may be explained on the grounds that most patients' symptoms were resolving spontaneously when they were admitted to the trial. However, as all but eight children were infected with at least one erythromycin-resistant enteropathogen in addition to C jejuni, the clinical failure of antibiotic therapy may have been due to erythromycin's inability to eliminate these organisms.

A hospital outbreak of multiresistant Salmonella typhimurium belonging to phage type 193

An outbreak of nosocomial infection was caused by strains of Salmonella typhimurium phage type 193 that were resistant to ampicillin, chloramphenicol, neomycin-kanamycin, streptomycin, spectinomycin, sulfonamides, tetracyclines, trimethoprim, and nalidixic acid. Resistances to drugs other than nalidixic acid were specified by plasmids, and, on the basis of phage typing and plasmid characterization studies, the multiresistant phage type 193 strains were determined to be clonal. In a two-year period, 488 patients infected with these bacteria were identified. An investigation in a pediatric surgical ward, where the outbreak was particularly severe, showed that patients exposed to antibiotics were more likely to be colonized with the epidemic strain and that young debilitated patients were more liable to show clinical signs of infection. Epidemiologic studies suggested that cross infection via the hands of the ward staff was the likely means of propagation of the epidemic.

Failure to detect conventional enterotoxins in classical enteropathogenic (serotyped) Escherichia coli strains of proven pathogenicity

Ammonium sulfate-precipitated supernatants of classical enteropathogenic Escherichia coli strains were negative when investigated for enterotoxin production in rabbit ligated ileal loops, rabbit skin vascular permeability factor tests, suckling mice, and Y-1 adrenal cells. They also failed to stimulate guanylate cyclase activity in homogenates of rabbit, rat, and infant mouse intestines. Furthermore, DNA from enteropathogenic E. coli lacked sequences that encode heat-labile and heat-stable enterotoxins. These studies fail to show conventional enterotoxin synthesis by classical enteropathogenic E. coli.

Detection of pneumococci in the upper respiratory tract: comparison of media and culture techniques

A survey for carriers of Streptococcus pneumoniae among 573 children was undertaken. Throat and nasopharyngeal cultures from 303 children disclosed pneumococci in 89 (29.4%) nasopharyngeal swabs but in only 8 (2.6%) throat specimens. Carriers of pneumococci were detected with frequencies of 84.4 and 79.9% on aerobically incubated 5% horse blood agar with a 5-micrograms optochin disk and on 5% horse blood agar with 5 mg of gentamicin per liter, respectively. Anaerobically incubated 5% horse blood agar yield pneumococci from 78.2% of 179 carriers. We consider the examination of nasopharyngeal secretions for optochin-susceptible alpha-hemolytic bacteria to be the most sensitive and convenient means for detecting carriers of pneumococci.

Magnitude, kinetics, and duration of vibriocidal antibody responses in North Americans after ingestion of Vibrio cholerae

Sera from 147 volunteers were examined for vibriocidal antibody by the tube-dilution method before and after they ingested 10(3)-10(6) Vibrio cholerae organisms. Titers increased significantly after challenge in 97% of 110 persons who excreted V. cholerae. In 12% of the bacteriologically confirmed infections, the titer increased significantly only to the homologous (Inaba or Ogawa) serotype. Levels of vibriocidal antibody decreased substantially between one and six months after challenge, but they usually remained elevated over base-line values. Levels of antitoxin to V. cholerae were measured by IgG enzyme-linked immunosorbent assay (ELISA), rabbit skin permeability tests, and adrenal cell techniques in 73 pairs of prechallenge and postchallenge sera. Results of the vibriocidal assay agreed most closely with those of the ELISA, the most sensitive antitoxin assay, in serologically detecting clinical and subclinical infections. Recent infection could be accurately serodiagnosed by levels of vibriocidal antibody and antitoxin (by IgG ELISA) in a convalescent-phase serum.

Lactobacillus prophylaxis for diarrhea due to enterotoxigenic Escherichia coli

In vitro and animal experiments indicated that lactobacilli might prevent Escherichia coli from colonizing the intestine and may produce substances counteracting enterotoxin. Lactinex, a commercial preparation of dried Lactobacillus acidophilus and L. bulgaricus, is marketed for uncomplicated diarrhea. Preliminary experiments in nonfasting volunteers indicated that lactobacilli in this preparation colonized the small intestine for up to 6 h. To evaluate the protective efficacy of Lactinex, a double-blind randomized study was carried out in which 48 volunteers (23 receiving Lactinex and 25 receiving placebos) were challenged with E. coli strains that produced heat-stable or heat-labile enterotoxins or both. No significant differences between the two groups were noted with respect to attack rate, incubation period, duration of diarrhea, volume and number of liquid stools, and coproculture yields. These data suggest that this lactobacillus preparations does not prevent or alter the course of enterotoxigenic E. coli diarrhea in adults. Lack of efficacy occurred despite efforts to maximize small bowel colonization, including administration of Lactinex in milk and in a 6-hour-interval regimen during 36 h before and 96 h after challenge.

The fate of ingested lactobacilli in the proximal small intestine

A freeze-dried commercial preparation of Lactobacillus acidophilus and Lactobacillus bulgaricus (Lactinex) dissolved in skim milk was ingested by four nonfasting and seven fasting informed community volunteers in the Isolation Ward of the Center for Vaccine Development. Samples of jejunal fluid withdrawn from the volunteers at varying intervals were cultured for lactobacilli on a selective medium. Quantitative counts varied considerably amongst the individuals studied and also in the same person examined on two consecutive occasions. In general, however, it was shown that lactobacilli entered the small intestine and persisted in elevated numbers for about 3 h in fasting subjects and for up to 6 h in nonfasting individuals. By 4 h, counts in fasting volunteers had returned to base-line levels. Although both Lactobacillus species in Lactinex entered the intestine in approximately equal numbers, L. acidophilus was recovered from more samples and in slightly greater number than L. bulgaricus.

Effect of chlorpromazine on intestinal secretion mediated by Escherichia coli heat-stable enterotoxin and 8-Br-cyclic GMP in infant mice

Subcutaneously administered chlorpromazine reduced intestinal fluid accumulation induced by Escherichia coli heat-stable enterotoxin in infant mice. The antisecretory effect of chlorpromazine, although dose related, was, even with high doses, less than that observed with respect to cholera toxin. Whereas 100 micrograms chlorpromazine abolished cholera toxin-induced intestinal secretion almost completely, 500 microgram chlorpromazine (equivalent to 200 microgram/g body wt) lowered secretion induced by heat-stable enterotoxin by only 41%. The effect of chlorpromazine on intestinal secretion was quantitatively similar regardless of whether heat-stable enterotoxin or the cyclic GMP analogue, 8-Br-cyclic GMP, was the secretagogue. This finding, which suggested that the inhibitory effect of chlorpromazine on heat-stable enterotoxin was independent of guanylate cyclase, was confirmed by assaying this enzyme in intestinal homogenates from mice that had been inoculated with chlorpromazine, and also in experiments in which chlorpromazine was added to guanylate cyclase assay mixtures in vitro. Caution is advised before chlorpromazine is routinely adopted for the treatment of all syndromes of watery diarrhea.

Summer diarrhoea in African infants and children

Of 70 black South African infants and children with acute summer diarrhoea, 30 (43%) were infected with enteropathogenic serogroups of Escherichia coli (EPEC), 13 (19%) with enterotoxigenic Gram-negative bacilli, 12 (17%) with Salmonella sp., 6 (9%) with Shigella sp., and 3 (4%) with rotaviruses. 13 (19%) patients were infected simultaneously with more than one enteropathogen, and no pathogen was detected in 22 (31%). In addition, 6 (15%) of 41 unselected patients were excreting Campylobacter fetus. Of 30 age-matched controls drawn from the same population, 5 (17%) were infected with EPEC serotypes, and 1 each with Salmonella sp. and rotavirus. This study stresses the polymicrobial nature of paediatric diarrhoea in a developing community and shows the continued importance of EPEC in this setting.

Mechanism of action of Yersinia enterocolitica enterotoxin

Enterotoxin derived from three clinical isolates of Yersinia enterocolitica was compared with the heat-stable enterotoxin of Escherichia coli. Both toxins were biologically active in infant mice examined at 2 h and in ligated rabbit ileal loops at 6 h. Neither substance, however, produced changes in ligated ileal loops at 18 h or in Chinese hamster ovary or Y1 adrenal tissue cultures. In addition, both Y. enterocolitica enterotoxin concentrated approximately 20 times by ammonium sulfate precipitation and ultrafiltration and a similarly prepared sample of E. coli heat-stable enterotoxin stimulated the activity of guanylate cyclase but not that of adenylate cyclase in infant mouse intestine. These findings suggest that the role of enterotoxin in the pathogenesis of intestinal Y.enterocolitica infection may be similar to that of heat-stable enterotoxin in E. coli diarrhea.

Antimicrobial susceptibility testing of pneumococci: determination of Kirby-Bauer breakpoints for penicillin G, erythromycin, clindamycin, tetracycline, chloramphenicol, and rifampin

Antimicrobial susceptibility testing of pneumococci is now essential to monitor for the presence of resistance to agents such as the penicillins, macrolides, lincomycins, chloramphenicol, and tetracycline. In this study, clinical isolates of a selection of resistant South African strains were tested for antimicrobial susceptibility by minimal inhibitory concentration (MIC) determination and by a modified Kirby-Bauer disk diffusion technique, using Mueller-Hinton medium supplemented with 5% horse blood. Disk diffusion breakpoints were determined for penicillin G, erythromycin, clindamycin, tetracycline, chloramphenicol, and rifampin. Reliable results were obtained on disk diffusion for all these agents except for penicillin G. With 6-mug penicillin G disks, zones of strains with intermediate penicillin susceptibility overlapped those of sensitive and resistant strains. With 5-mug methicillin disks, clearer separation of strains based on susceptibility to penicillin G occurred. Strains with zones of <35 mm around penicillin G disks and <25 mm around methicillin disks should have penicillin G MICs determined to confirm their resistance to penicillin G. In view of the potential for pneumococci to be resistant to the agents used in this study, antimicrobial susceptibility of all clinically significant isolates should be determined.

Yersinia enterocolitica biotype 1 in South Africa

Two clinical isolates of Yersinia enterocolitica biotype 1, positive for indole and lecithinase production, xylose and salicin fermentation, and aesculin hydrolysis, are reported. One strain, serotype 0 6, was recovered from a healthy food handler, while the other, serotype 0 5, was associated with a diarrhoeal illness in a 1-year-old child. This is the first report of either of these serotypes from human material in South Africa.

Etiology of infantile enteritis in South Africa

Infantile enteritis constitutes a major health problem in developing countries. Several investigations into the etiology of this condition among various South African populations have been undertaken during the past few years. Recent studies of black urban infants have revealed that salmonellae, shigellae, enterotoxigenic Enterobacteriaceae, and rotaviruses play a relatively minor role in infantile enteritis. On the other hand, all studies, including a number performed several years ago, have demonstrated that Escherichia coli strains belonging to traditional enteropathogenic serotypes are more frequently recovered from patients with diarrhea than from matched control subjects. These bacteria are particularly prominent during the summer months at the height of the annual gastroenteritis epidemic. Recent studies have indicated the importance of Campylobacter fetus in infantile enteritis, especially in children younger than nine months. The data point to a complex multiple etiology of infantile enteritis in South Africa, which will require clarification before specific preventive measures can be instituted.

Antibiotic susceptibility of Neisseria gonorrhoeae isolated in Johannesburg

The susceptibility of 175 clinical isolates of Neisseria gonorrhoeae to penicillin G, ampicillin, erythromycin, tetracycline and spectinomycin was determined. Almost one-fifth of strains were relatively resistant to penicillin G. Relative resistance to penicillin G was significantly correlated with resistance to ampicillin, tetracycline and erythromycin, but not with resistance to spectinomycin. No penicillinase-producing or spectinomycin-resistant strains were encountered in this study. The relevance of the findings to the management of gonococcal urethritis is discussed.

Emergence of multiply resistant pneumococci

Multiple antimicrobial resistance in pneumococci was detected in Johannesburg in July, 1977, and prompted an investigation of the prevalence of resistant strains in two hospitals. Carriers of Types 6A and 19A penicillin-resistant pneumococci, resistant to antibiotic concentrations ranging between 0.12 and 4 microgram per milliliter were found in 29 per cent of 543 pediatric patients and 2 per cent of 434 hospital staff members. Multiply resistant Type 19A strains, resistant to beta-lactam antibiotics, erythromycin, clindamycin, tetracycline and chloramphenicol, were isolated from 128 carriers, and were responsible for bacteremia in four patients. Isolates from 40 other carriers were resistant to penicillin alone or to penicillin and chloramphenicol or to penicillin, chloramphenicol and tetracycline. Pneumococci can be screened for penicillin resistance with a modified Kirby--Bauer technic; the strains with zones of less than 35 mm around 6-microgram penicillin disks or less than 25 mm around 5-microgram methicillin disks should be tested for sensitivity to penicillin by measurements of minimum inhibitory concentration.

Escherichia coli gastro-enteritis in adults

Two cases of gastro-enteritis in South African adults are reported. Stool cultures performed on both patients yielded pure growths Escherichia coli. One isolate was shown to have invasive ability, and the other to produce heatstable enterotoxin. Neither strain belonged to a 'classic' enteropathogenic serotype associated with infantile gastro-enteritis.

Rotavirus and winter gastro-enteritis in White South African infants

In a microbiological investigation of winter gastro-enteritis in 23 White children rotaviruses were found in 14 (61%) and parvovirus-like particles in 1 (4%). Bacteriological examination of stools from 11 of these patients yielded one isolate of Salmonella eastbourne, but no enterotoxigenic or invasive bacteria were found. Rotavirus appears to be the main aetiological agent of acute winter gastro-enteritis in White infants.

Pathogenic mechanisms of a non-agglutinable Vibrio cholerae strain: demonstration of invasive and enterotoxigenic properties

A non-agglutinable Vibrio cholerae strain isolated from the blood of a child with kwashiorkor and fever was shown to have the potential to invade as well as to produce a toxin resembling cholera toxin.

A microbiological study of gastro-enteritis in Black infants

A survey was carried out to determine the aetiological agents responsible for acute gastro-enteritis in Black infants. Bacteriological examination of 191 patients and 178 age-matched controls showed that salmonellae and shigellae together accounted for only 15% of cases and were harboured by almost 7% of controls. Seropositive 'enteropathogenic' Escherichia coli was an important cause of diarrhoea, especially in the summer months. Enterotoxin-producing E. coli was not a significant cause of gastro-enteritis in this study. This study suggests that routine serotyping of E. coli from infantile gastro-enteritis should be continued until the relationship of serotype and enterotoxigenicity with enteropathogenicity has been established. Rotaviruses were detected by electron microscopy in 14,4% of children. The clinical features associated with the different infectious agents are described.

Urethritis in white men--a microbiological appraisal

The incidence of infections with Neisseria gonorrhoeae, Chlamydia trachomatis and mycoplasmas has been assessed in White men with urethritis who were attending a Johannesburg clinic for sexually transmitted diseases. Infection with N. gonorrhoeae was demonstrated in 54% of 212 patients with urethritis. C. trachomatis was isolated from 37% of patients with non-gonococcal urethritis, and from 25% with proven gonococcal infection. Although T-mycoplasmas were isolated significantly more often from patients with urethritis (41%) than from controls (23%) (P less than 0,05) their role in the aetiology of urethritis is uncertain.

Thrombocytopenia in trypanosomiasis

In all of four patients with African trypanosomiasis, thrombocytopenia was present on admission to hospital or developed during the course of the illness. One patient with severe thrombocytopenia died following gastrointestinal haemorrhage shortly after admission to hospital. Kinetic studies in the other three patients showed marked pooling of platelets in the spleen in all, but severe shortening of platelet life-span in only one. Evidence of disseminated intravascular coagulation (DIC) was found in 3 patients, 2 of whom received heparin therapy. These findings provide evidence that thrombocytopenia is a feature of African trypanosomiasis and is due mainly to hypertrophy of the reticuloendothelial system which accompanies the infection. In some patients immune damage to platelets or platelet consumption as part of DIC may be an additional factor contributing to the thrombocytopenia.