Performance of the chromID Salmonella Elite chromogenic agar in comparison with CHROMagar™ Salmonella, Oxoid™ Brilliance™ Salmonella and Hektoen agars for the isolation of Salmonella from stool specimens.

chromID™ Salmonella Elite is compared with 3 culture media commonly used for Salmonella isolation from stool specimens. As results were equivalent to other chromogenic media (100% sensitivity, 98% specificity), only financial arguments should guide the choice for a medium with respect to another.

Variations in catheter-related bloodstream infections rates based on local practices.

BACKGROUND: Catheter-related bloodstream infection (CRBSI) surveillance serves as a quality improvement measure that is often used to assess performance. We reviewed the total number of microbiological samples collected in three Belgian intensive care units (ICU) in 2009-2010, and we described variations in CRBSI rates based on two factors: microbiological documentation rate and CRBSI definition which includes clinical criterion for coagulase-negative Staphylococcus (CNS) episode. FINDINGS: CRBSI rates were 2.95, 1.13 and 1.26 per 1,000 estimated catheter-days in ICUs A, B and C, respectively. ICU B cultured fewer microbiological samples and reported the lowest CRBSI rate. ICU C had the highest documentation rate but was assisted by support available from the laboratory for processing single CNS positive blood cultures. With the exclusion of clinical criterion, CRBSI rates would be reduced by 19%, 45% and 0% in ICUs A, B and C, respectively. CONCLUSION: CRBSI rates may be biased by differences of blood culture sampling and CRBSI definition. These observations suggest that comparisons of CRBSI rates in different ICUs remain difficult to interpret without knowledge of local practices.

Comparison of the Microflex LT and Vitek MS systems for routine identification of bacteria by matrix-assisted laser desorption ionization-time of flight mass spectrometry.

This study compared the performance of three matrix-assisted laser desorption ionization-time of flight mass spectrometry systems: Microflex LT (Bruker Daltonics, Bremen, Germany), Vitek MS RUO (Axima Assurance-Saramis database; bioMérieux, Marcy l'Etoile, France), and Vitek MS IVD (bioMérieux). A total of 1,129 isolates, including 1,003 routine isolates, 73 anaerobes, and 53 bacterial enteropathogens, were tested on the Microflex LT and Axima Assurance devices. The spectra were analyzed using three databases: Biotyper (Bruker Daltonics), Saramis, and Vitek MS (bioMérieux). Among the routine isolates requiring identification to the species level (n = 986), 92.7% and 93.2% were correctly identified by the Biotyper and Vitek MS databases, respectively. The Vitek MS database is more specific for the identification of Streptococcus viridans. For the anaerobes, the Biotyper database often identified Fusobacterium isolates to only the genus level, which is of low clinical significance, whereas 20% of the Bacteroides species were not identified or were misidentified by the Vitek MS database. For the enteropathogens, the poor discrimination between Escherichia coli and Shigella explains the high proportion of unidentified organisms. In contrast to the Biotyper database, the Vitek MS database properly discriminated all of the Salmonella entrica serovar Typhi isolates (n = 5). The performance of the Saramis database was globally poorer. In conclusion, for routine procedures, the Microflex LT and Vitek-MS systems are equally good choices in terms of analytical efficiency. Other factors, including price, work flow, and lab activity, will affect the choice of a system.

Antimicrobial susceptibility of Streptococcus pneumoniae isolates from vaccinated and non-vaccinated patients with a clinically confirmed diagnosis of community-acquired pneumonia in Belgium.

We assessed the in vitro susceptibility of Streptococcus pneumoniae isolates from patients with confirmed community-acquired pneumonia (CAP) to ß-lactams, macrolides and fluoroquinolones and the association of non-susceptibility and resistance with serotypes/serogroups (STs/SGs), patient's risk factors and vaccination status. Samples (blood or lower respiratory tract) were obtained in 2007-2009 from 249 patients (from seven hospitals in Belgium) with a clinical and radiological diagnosis of CAP [median age 61 years (11.6% aged <5 years); 85% without previous antibiotic therapy; 86% adults with level II Niederman's severity score]. MIC determination (EUCAST breakpoints) showed for: (i) amoxicillin, 6% non-susceptible; cefuroxime (oral), 6.8% resistant; (ii) macrolides: 24.9% erythromycin-resistant [93.5% erm(B)-positive] but 98.4% telithromycin-susceptible; and (iii) levofloxacin and moxifloxacin, all susceptible. Amongst SGs: ST14, all resistant to macrolides and most intermediate to ß-lactams; SG19 (>94% ST19A), 73.5% resistant to macrolides and 18-21% intermediate to ß-lactams; and SG6, 33% resistant to clarithromycin. Apparent vaccine failures: 3/17 for 7-valent vaccine (children; ST6B, 23F); 16/29 for 23-valent vaccine (adults ST3, 7F, 12F, 14, 19A, 22F, 23F, 33F). Isolates from nursing home residents, hospitalised patients and patients with non-respiratory co-morbidities showed increased MICs for amoxicillin, all ß-lactams, and ß-lactams and macrolides, respectively. Regarding antibiotic susceptibilities: (i) amoxicillin is still useful for empirical therapy but with a high daily dose; (ii) cefuroxime axetil and macrolides (but not telithromycin) are inappropriate for empirical therapy; and (iii) moxifloxacin and levofloxacin are the next 'best empirical choice' (no resistant isolates) but levofloxacin will require 500 mg twice-daily dosing for effective coverage.

Incidence and virulence determinants of verocytotoxin-producing Escherichia coli infections in the Brussels-Capital Region, Belgium, in 2008-2010.

The incidence of verocytotoxin-producing Escherichia coli (VTEC) was investigated by PCR in all human stools from Universitair Ziekenhuis Brussel (UZB) and in selected stools from six other hospital laboratories in the Brussels-Capital Region, Belgium, collected between April 2008 and October 2010. The stools selected to be included in this study were those from patients with hemolytic-uremic syndrome (HUS), patients with a history of bloody diarrhea, patients linked to clusters of diarrhea, children up to the age of 6 years, and stools containing macroscopic blood. Verocytotoxin genes (vtx) were detected significantly more frequently in stools from patients with the selected conditions (2.04%) than in unselected stools from UZB (1.20%) (P = 0.001). VTEC was detected most frequently in patients with HUS (35.3%), a history of bloody diarrhea (5.15%), or stools containing macroscopic blood (1.85%). Stools from patients up to the age of 17 years were significantly more frequently vtx positive than those from adult patients between the ages of 18 and 65 years (P = 0.022). Although stools from patients older than 65 years were also more frequently positive for vtx than those from patients between 18 and 65 years, this trend was not significant. VTEC was isolated from 140 (67.9%) vtx-positive stools. One sample yielded two different serotypes; thus, 141 isolates could be characterized. Sixty different O:H serotypes harboring 85 different virulence profiles were identified. Serotypes O157:H7/H- (n = 34), O26:H11/H- (n = 21), O63:H6 (n = 8), O111:H8/H- (n = 7), and O146:H21/H- (n = 6) accounted for 53.9% of isolates. All O157 isolates carried vtx2, eae, and a complete O island 122 (COI-122); 15 also carried vtx1. Non-O157 isolates (n = 107), however, accounted for the bulk (75.9%) of isolates. Fifty-nine (55.1%) isolates were positive for vtx1, 36 (33.6%) were positive for vtx2, and 12 (11.2%) carried both vtx1 and vtx2. Pulsed-field gel electrophoresis revealed wide genetic diversity; however, small clusters of O157, O26, and O63:H6 VTEC that could have been part of unidentified outbreaks were identified. Antimicrobial resistance was observed in 63 (44.7%) isolates, and 34 (24.1%) showed multidrug resistance. Our data show that VTEC infections were not limited to patients with HUS or bloody diarrhea. Clinical laboratories should, therefore, screen all stools for O157 and non-O157 VTEC using selective media and a method for detecting verocytotoxins or vtx genes.

Performance of individual Helicobacter pylori antigens in the immunoblot-based detection of H. pylori infection.

To develop a specific line blot (LB) for supporting ELISA-based serodiagnosis of Helicobacter pylori infection, individual native/recombinant H. pylori antigens were evaluated with respect to their reactivity with both serum IgG and IgA from 156 dyspeptic screening patients (67% H. pylori positive). Of 13 antigens, HP0175, p17, and p19 revealed highest positive likelihood ratios for H. pylori-specific IgG (> 5.0) and were selected as LB substrates, in addition to the established virulence markers VacA and CagA. For validation, the LB was compared to a commercial whole-cell-lysate-based ELISA by parallel (re-)analysis of 156 screening sera, 22 sera from diabetes mellitus patients and 15 sera from follow-up patients after H. pylori eradication. In screening patients, the combined use of IgG ELISA and LB revealed a sensitivity, specificity, and accuracy of 94%, 81%, and 90%, respectively, whereas IgG ELISA alone exhibited a low specificity of 75%. In diabetic and follow-up patients, IgA ELISA exhibited high accuracy of 89% and 93%, respectively, whereas IgG detection was unreliable (accuracy < 80%). In conclusion, using HP0175, p17, p19, CagA, and VacA as LB substrates significantly improves the specificity of anti-H. pylori IgG analysis, providing a reliable tool for (1) confirmation/refutation of ELISA-based screening results and (2) assessment of the CagA/VacA status.

New identification characteristics of methicillin-resistant Staphylococcus aureus on chromogenic culture media.

Methicillin-Resistant Staphylococcus aureus (MRSA) is one of the most common and important causes of nosocomial infections. Rapid detection of this pathogen is important for conducting good and swift infection control. This prospective study evaluates two chromogenic media for the detection of MRSA. New colony characteristics were noticed during this evaluation: (i) a yellow/golden colouration on a pipette after streaking the colonies of the chromogenic culture could eventually be used as a supplementary identification test to identify the MRSA strains, and (ii) some MRSA strains do not metabolise the chromogens and therefore are not coloured on chromogenic agars. However, they have a typical yellow/golden colony aspect usually observed amongst S. aureus.

In vivo development of antimicrobial resistance in Pseudomonas aeruginosa strains isolated from the lower respiratory tract of Intensive Care Unit patients with nosocomial pneumonia and receiving antipseudomonal therapy.

Pseudomonas aeruginosa causes severe nosocomial pneumonia in Intensive Care Unit (ICU) patients, with an increased prevalence of multiresistant strains. We examined the impact of the use of antipseudomonal antibiotic(s) on the susceptibility of P. aeruginosa isolated from ICU patients with clinically suspected hospital-acquired pneumonia collected in five teaching hospitals (110 non-duplicate initial isolates; 62 clonal pairs of initial and last isolates during treatment). Minimum inhibitory concentrations (MICs) were determined for amikacin, ciprofloxacin, meropenem, piperacillin/tazobactam (TZP), cefepime and ceftazidime (used in therapy) as well as five reporter antibiotics (aztreonam, colistin, gentamicin, piperacillin and ticarcillin) using Clinical and Laboratory Standards Institute (CLSI) methodology. Susceptibility was assessed according to European Committee on Antimicrobial Susceptibility Testing (EUCAST) and CLSI breakpoints. Resistance rates prior to treatment exceeded 25% for cefepime, ceftazidime, piperacillin, ticarcillin and aztreonam (EUCAST and CLSI) and for gentamicin, TZP and colistin (EUCAST only). The highest rates of cross-resistance were noted for ceftazidime and cefepime and the lowest rate for amikacin. Mean MIC values were systematically higher in isolates from patients previously exposed (1 month) to the corresponding antibiotic. For clonal pairs, a systematic increase in MIC between initial and last isolates (significant for amikacin, cefepime, meropenem and TZP) was noted. There was a significant correlation between the use of antibiotics (adjusted for respective proportional use of each drug) and loss of susceptibility at the population level when using EUCAST breakpoints. The high level of resistance of P. aeruginosa in ICU patients with nosocomial pneumonia as well as its further increase during treatment severely narrows the already limited therapeutic options. Further observational studies and the development of early diagnosis for resistant isolates are warranted.

Vitek 2 ANC card versus BBL Crystal Anaerobe and RapID ANA II for identification of clinical anaerobic bacteria.

The Vitek 2 Anaerobe and Corynebacterium Identification Card (ANC) was recently evaluated in a multicentre study. In the present work, this system was compared with the BBL Crystal Anaerobe and RapID ANA II panels. These kits were tested using 196 strains of anaerobes that had been previously identified by gas-liquid chromatography. Identification to the species or to the genus level was 75.0%, 81.1% and 70.9% for Crystal, RapID and Vitek, respectively. Vitek ANC failed to provide any identification in 20.4% of the strains, but it had fewer misidentifications than RapID. The confidence factors provided on the results report of each kit were not always correlated with a lower risk of major errors, with the exception of Vitek 2 in which a confidence factor higher than 0.86 excluded the risk of misidentification in more than 87% of isolates. The lower rate of identification by the Vitek and Crystal panels is mostly due the lower ability of these systems to identify the Clostridia. Overall, the three panels are comparable but need improvement to a better accuracy.

High incidence of invasive group B streptococcal infections in HIV-exposed uninfected infants.

OBJECTIVES: The occurrence of an unusual number of group B streptococcal (GBS) infections in HIV-exposed uninfected (HEU) infants who were followed in our center prompted this study. The objective of this study was to describe and compare the incidence and clinical presentation of GBS infections in infants who were born to HIV-infected and -uninfected mothers. METHODS: All cases of invasive GBS infections in infants who were born between 2001 and 2008 were identified from the database of HEU infants and from the microbiology laboratory records. The medical charts of all infants with GBS infection were reviewed. RESULTS: GBS invasive infections were described for 5 (1.55%) infants who were born to 322 HIV-infected mothers who delivered in our center. The incidence of GBS infections during the same period was 16 (0.08%) of 20 158 infants who were born to HIV-uninfected mothers. One HEU infant presented a recurrent infection 28 days after completion of treatment for the first episode. Late-onset infection was more frequent in HEU infants (5 of 6 vs 2 of 16 episodes in the control population). The diseases were also more severe in HEU infants with 5 of 6 sepsis or sepsis shock in HEU infants versus 10 of 16 in control subjects, and most HEU infants had leukopenia at onset of infection. CONCLUSIONS: The incidence of GBS infection was significantly higher in HEU infants than in infants who were born to HIV-uninfected mothers. These episodes of GBS sepsis in HEU infants were mostly of late onset and more severe than in the control population, suggesting an increased susceptibility of HEU infants to GBS infection.

Heparin-binding, hemagglutinin-specific IFN-gamma synthesis at the site of infection during active tuberculosis in humans.

RATIONALE: Tuberculosis (TB) remains a major cause of mortality. A better understanding of the immune responses to mycobacterial antigens may be helpful to develop improved vaccines and diagnostics. OBJECTIVES: The mycobacterial antigen heparin-binding hemagglutinin (HBHA) induces strong IFN-γ responses by circulating lymphocytes from subjects latently infected with Mycobacterium tuberculosis, and low responses associated with CD4(+) regulatory T (Treg) cells in patients with TB. Here, we investigated HBHA-specific IFN-γ responses at the site of the TB disease. METHODS: Bronchoalveolar lavages, pleural fluids, and blood were prospectively collected from 61 patients with a possible diagnosis of pulmonary or pleural TB. HBHA-specific IFN-γ production was analyzed by flow cytometry and ELISA. The suppressive effect of pleural Treg cells was investigated by depletion experiments. MEASUREMENTS AND MAIN RESULTS: The percentages of HBHA-induced IFN-γ(+) alveolar and pleural lymphocytes were higher for pulmonary (P < 0.0001) and for pleural (P < 0.01) TB than for non-TB controls. Local CD4(+) and CD8(+) T cells produced the HBHA-specific IFN-γ. This local secretion was not suppressed by Treg lymphocytes, contrasting with previously reported data on circulating lymphocytes. CONCLUSIONS: Patients with TB display differential effector and regulatory T-cell responses to HBHA in local and circulating lymphocytes with a predominant effector CD4(+) and CD8(+) response locally, compared with a predominant Treg response among circulating lymphocytes. These findings may be helpful for the design of new vaccines against TB, and the detection of HBHA-specific T cells at the site of the infection may be a promising tool for the rapid diagnosis of active TB.

A prospective evaluation of selection criteria for the use of Nucleic Acid Amplification Techniques to test for the presence of Mycobacterium tuberculosis in respiratory tract samples.

BACKGROUND: Although Mycobacterium tuberculosis (MTB) can be detected rapidly by means of Nucleic Acid Amplification Techniques (NAT), these NAT tests are expensive and therefore are not used in routine practice or as a screening tool. METHODS: Although it is generally accepted that clinical and radiological data are important markers for deciding whether to test for MTB using NAT, the optimal combination of markers has not been determined. A prospective study was performed to evaluate NAT using different combinations of clinical, laboratory and radiographic selection criteria. RESULTS: The sensitivity of NAT for detecting MTB in patients with smear negative for acid-fast bacilli was two times higher in patients with radiographic abnormalities and negative routine bacterial culture for respiratory pathogens than those with absence of radiographic abnormalities or positive routine bacterial culture (77.8% vs. 33.3%; p<0.001). Furthermore, a difference of almost 40% can be observed between the positive predictive values of both groups (87.5% vs. 50.0%; p<0.001). CONCLUSIONS: Using a combination of clinical, laboratory and radiographic criteria, it is possible to identify patients in whom the NAT for MTB has reasonable sensitivity and specificity. Using these selection criteria should reduce costs associated with the inappropriate use of NAT tests.

Diagnostic value of five commercial tests for the rapid diagnosis of Clostridium difficile-associated disease.

We compared five commercial immunoassays (Biostar OIA CdTOX AB, ImmunoCard Toxins A&B - Meridian, Xpect C. difficile toxin A/B -Remel, C. difficile toxin A test- Oxoid, and TOX A/B QUIK CHEK- Techlab) which allow a rapid diagnosis of C. difficile associated disease. The tests were performed directly on patient's stool specimen submitted for routine investigation of C. difficile infection from two University Hospitals in Brussels. The cell cytotoxicity assay was considered as the gold standard. Of the 100 stool specimens included in the study 23 were positive for C. difficile toxin. The sensitivity, specificity, positive and negative predictive values were respectively, 95.7%, 100%, 100% and 98.7% for TOX A/B QUIK CHEKTM, 91.3%, 100%, 100% and 97.5% for ImmunoCard Toxins A&B and for Xpect C. difficile toxin A/B, 87%, 100%, 100% and 96.3% for OIA CdTOX AB and 87%, 98.7%, 97.2% and 96.3% for C. difficile toxin A test. The differences were not statistically significant (p > 0.05). These data suggest that the tested immunoassays are acceptable for rapid detection of C. difficile toxin.

Treatment of Dientamoeba fragilis infection with paromomycin.

Treatment with paromomycin (25-35 mg/kg/d for 7 days) was evaluated prospectively in 15 children with Dientamoeba fragilis infection after 1-month follow-up. At the end of the study, parasitologic effectiveness and clinical improvement were observed in 12/15 (80%) and 13/15 (87%) patients, respectively. Paromomycin appears to be an effective drug for treatment of D. fragilis infection in children.

Antimicrobial susceptibility of clinical isolates of non-jejuni/coli campylobacters and arcobacters from Belgium.

OBJECTIVES: To determine the susceptibility of non-jejuni/coli campylobacters and arcobacters isolated from diarrhoeal stool specimens in Belgium. METHODS: The MICs were determined using Etest for six antimicrobial agents including ampicillin, erythromycin, nalidixic acid, ciprofloxacin, gentamicin and tetracycline for the most frequently isolated non-jejuni/coli campylobacter and arcobacter strains in two University Hospital laboratories between 1995 and 2005. RESULTS: In total, 85 Campylobacter upsaliensis, 20 Campylobacter concisus, 11 Campylobacter fetus, 61 Arcobacter butzleri and 10 Arcobacter cryaerophilus isolates were tested. Most C. upsaliensis strains were susceptible to ampicillin (100%), gentamicin (100%), ciprofloxacin (94.1%) and tetracycline (100%), whereas 11.8 and 12.9% were resistant to nalidixic acid and erythromycin, respectively. For A. butzleri, 78.7% of isolates were susceptible to ampicillin and erythromycin. Most A. butzleri isolates were susceptible to ciprofloxacin (96.7%), nalidixic acid (82.0%), gentamicin (100%) and tetracycline (100%). All C. concisus strains were fully susceptible to ampicillin and tetracycline, but 5% of them were resistant to gentamicin, ciprofloxacin and erythromycin. Nearly all C. fetus and A. cryaerophilus strains were susceptible to erythromycin but the results should be interpreted with caution since only a small number of strains were tested. CONCLUSIONS: Fluoroquinolones should be considered in the treatment of severe C. upsaliensis and A. butzleri infection. When clinically indicated, erythromycin remains the first choice for the treatment of intestinal campylobacteriosis caused by C. concisus and C. fetus.

Clinical and microbiological features of dientamoebiasis in patients suspected of suffering from a parasitic gastrointestinal illness: a comparison of Dientamoeba fragilis and Giardia lamblia infections.

OBJECTIVES: To describe the clinical and microbiological features of Dientamoeba fragilis and Giardia lamblia infected patients, and to analyze the genetic variation of D. fragilis strains. METHODS: For a period of two years, all stool samples collected from patients suspected of having a parasitic gastrointestinal infection were examined according to our specific triple feces test (TFT) protocol. A retrospective case-control study was performed on D. fragilis and G. lamblia infected patients. Furthermore, PCR and genotyping by restriction fragment length polymorphism (RFLP) were performed upon the former. RESULTS: D. fragilis (6.3%) and G. lamblia (7.1%) were the most common pathogenic protozoa isolated out of 448 patients studied. Symptoms most frequently encountered with D. fragilis and G. lamblia infection were abdominal pain (69.2% and 72.4%, respectively) and diarrhea (61.5% and 79.3%, respectively). However, patients with D. fragilis infections suffered significantly less frequently from nausea and/or vomiting, anorexia and weight loss. After treatment, all D. fragilis and G. lamblia infected patients presenting a negative TFT follow-up also reported a complete resolution of their symptoms. Only genotype 1 could be detected in D. fragilis infected patients. CONCLUSIONS: D. fragilis and G. lamblia were the most frequently encountered parasites in our study population. Improved diagnostic tests are essential tools to study the prevalence and pathogenesis of D. fragilis.

Improvement of routine diagnosis of intestinal parasites with multiple sampling and SAF-fixative in the triple-faeces-test.

BACKGROUND AND STUDY AIM: To perform optimal laboratory diagnosis of intestinal parasites is demanding. Because intestinal parasites are intermittently shedded, examination of multiple stools is imperative. For reliable detection of vegetative stages of protozoa, fresh stools should be examined direct after production, or stools should be preserved in a fixative. These aspects in routine practice are often neglected with as a result lower sensitivity of the diagnostic procedure. With application of the Triple-Faeces-Test (TFT) protocol, where both multiple sampling and a SAF-fixative are included, these practical problems could be overcome. The aim of this study was to compare the recovery of intestinal parasites in faecal specimens using TFT protocol versus the conventional diagnostic method (ether-sedimentation of one fresh stool sample). METHODS: During a three years period, results obtained in routine practice with the TFT protocol were compared with results from examination of sediment obtained with the ethyl-acetate-sedimentation technique of one unpreserved faeces specimen. RESULTS: From 2776 patients, 28.1% were positive for one or more intestinal parasites after examination of the TFT test, compared to 10.3% positivity with the conventional method (P < 0.05). Pathogenic species and non pathogenic species were observed respectively 191 and 449 times with TFT and 105 and 152 times with conventional method (P < 0.05). CONCLUSIONS: The application of the Triple-Faeces-Test in routine clinical practice significantly increased recovery of intestinal parasitic infections.

Arcobacter species in humans.

During an 8-year study period, Arcobacter butzleri was the fourth most common Campylobacter-like organism isolated from 67,599 stool specimens. Our observations suggest that A. butzleri displays microbiologic and clinical features similar to those of Campylobacter jejuni; however, A. butzleri is more frequently associated with a persistent, watery diarrhea.

Hemolytic uremic syndrome in Belgium: incidence and association with verocytotoxin-producing Escherichia coli infection.

OBJECTIVE: To evaluate the incidence of hemolytic uremic syndrome (HUS) in Belgium and to determine the role of verocytotoxin-producing Escherichia coli O157:H7 and other serotypes (non-O157 VTEC). METHODS: Twenty-two centers, including the seven university hospitals, registered prospectively all cases of HUS; they collected clinical samples for isolation of VTEC strains and serum for detection of specific O-lipopolysaccharide antibodies. RESULTS: Forty-seven cases of HUS (including five incomplete cases) were recorded. Three cases were seen in non-residents. The incidence of complete HUS in Belgian residents was 4.3 cases/100 000 in children <5 years old, 1.8 cases/100 000 when all children <15 years were considered, and 0.42/100 000 when patients of all ages were taken into account. By combining bacteriologic and serologic results, evidence of VTEC infection was obtained in 64% of the patients, mainly but not exclusively in children with prodromal diarrhea. The 13 VTEC isolates belonged to serotypes O157:H7 (nine isolates), O26:H11, O121:H---, O145:H--- and O172:H--- (one each) and all produced VT2 (+VT2vh-a in three O157 strains) and were positive for the eaeA gene. CONCLUSIONS: The incidence rate found in this study and the high mortality and morbidity linked with this syndrome warrant further registration of pediatric and post-diarrheic adult HUS cases and also examination of stools for both O157 and non-O157 VTEC strains. For effective prevention of this disease, further study of the serotypes and accessory virulence factors associated with HUS is needed.