To evaluate intrahospital and interhospital clonal dissemination of extended-spectrum beta-lactamase (ESBL)-producing strains of Klebsiella pneumoniae.

Eight tertiary-care university hospitals and 16 regional hospitals in Taiwan.

Two hundred eleven confirmed ESBL-producing isolates of K. pneumoniae were collected from January 1998 to June 2000. The isolates were characterized by various typing methods, including antibiogram (9 antimicrobial agents), computer-based ribotyping, pulsed-field gel electrophoresis (PFGE), and isoelectric focusing of beta-lactamase.

Ribotyping identified 70 distinct ribogroups among 200 isolates evaluated. Forty-three of these ribogroups were unique. Eleven ribogroups, comprising 115 isolates, were detected in more than one hospital (interhospital dissemination), whereas 16 groups (42 isolates) were detected in more than one patient within a hospital (intrahospital dissemination). The combination of ribotyping and PFGE identified two large epidemic clones, which were called 691.5/PFGE-G and 595.7/PFGE-A. These epidemic clones were detected mainly in the hospitals located in the northern and central regions of Taiwan. However, variation of the profiles of antibiograms and isoelectric focusing was apparent within each clone. In addition, isolates with the same isoelectric focusing profile (isoelectric points 7.9, 8.2, and 8.4) and antibiogram (resistance to 9 compounds evaluated) were present among different molecular-typed clones.

Our results showed that clonal dissemination (both interhospital and intrahospital dissemination) is occurring in several regions of Taiwan. Rapid computer-based ribotyping associated with PFGE demonstrated multiple epidemic clones of ESBL-producing K. pneumoniae in Taiwan. The combination of phenotypic and molecular methods has proved useful to characterize these epidemic clones.

To investigate the clinical features and the epidemiology of Acinetobacter baumannii in Spanish hospitals.

Prospective multicenter cohort study.

Twenty-seven general hospitals and one paraplegic center in Spain.

All cases of A. baumannii colonization or infection detected by clinical samples during November 2000 were included. Isolates were identified using phenotypic and genotypic methods. The molecular relatedness of the isolates was assessed by pulsed-field gel electrophoresis.

Twenty-five (89%) of the hospitals had 221 cases (pooled rate in general hospitals, 0.39 case per 1,000 patient-days; range, 0 to 1.17). The rate was highest in intensive care units (ICUs). Only 3 cases were pediatric. The mean age of the patients in the general hospitals was 63 years; 69% had a chronic underlying disease and 80% had previously received antimicrobial treatment. Fifty-three percent of the patients had an infection (respiratory tract, 51%; surgical site, 16%; and urinary tract, 11%). Crude mortality was higher in infected than in colonized patients (27% vs 10%; relative risk, 1.56; 95% confidence interval, 1.2 to 2.0; P = .003). Molecular analysis disclosed 79 different clones. In most hospitals, a predominant epidemic clone coexisted with other sporadic clones. Imipenem resistance was present in 39% of the hospitals.

A. baumannii was present in most participating Spanish hospitals (particularly in ICUs) with different rates among them. The organisms mainly affected predisposed patients; half of them were only colonized. Epidemic and sporadic clones coexisted in many centers.

To describe the epidemiology of endemic multidrug-resistant Pseudomonas aeruginosa colonizations and infections in critically ill patients.

Prospective study on bacterial strain typing and retrospective cohort study of charts of patients in the intensive care unit (ICU).

Fifty-three patients with P. aeruginosa isolated from clinical cultures in 2001 were selected, divided into those with P. aeruginosa in vitro resistant to at least two classes of antibiotics (multidrug-resistant, n = 18) and those susceptible to all or resistant to only one antibiotic (susceptible, n = 35).

Risk factors for multidrug-resistant P. aeruginosa included maxillary sinusitis, long-dwelling central venous catheters, prolonged use of certain antibiotics, a high lung injury score, and prolonged mechanical ventilation and duration of stay. The frequency of colonization (approximately 50%) versus infection (ie, ventilator-associated pneumonia) did not differ between the groups. On amplified fragment-length polymorphism analysis, 64% of the multidrug-resistant strains had been potentially transmitted via cross-colonization and 36% had probably originated endogenously. ICU mortality was 22% in the multidrug-resistant group and 23% in the susceptible group, although the duration of mechanical ventilation was longer in the former.

Patients with sinusitis who stayed in the ICU longer, were ventilated longer because of acute lung injury, received antibiotics for longer durations, and had long-dwelling central venous catheters ran an elevated risk of acquiring multidrug-resistant P. aeruginosa. These patients did not have a higher mortality than patients with susceptible P. aeruginosa. Prevention of the emergence of multidrug-resistant strains requires changes in infection control measures and antibiotic policies in our ICU.

To identify risk factors for the respiratory acquisition of extended-spectrum beta-lactamase (ESBL)-producing Klebsiella pneumoniae among patients admitted to a neurosurgical intensive care unit (NSICU) and to modify them without changing general infection control measures.

Nested case-control and intervention study.

A 1,200-bed, tertiary-care teaching hospital with a 17-bed NSICU.

Sputa of all patients admitted to the NSICU were cultured weekly during the study. From October 2002 through February 2003, 29 case-patients from whose sputum ESBL-producing K. pneumoniae was isolated were detected and 59 controls-patients were randomly selected among patients without any positive isolate of ESBL-producing K. pneumoniae. After analyzing the risk factors, we intervened to modify them and compared the acquisition rate of ESBL-producing K. pneumoniae before (October 2002 to February 2003) and after (April to August 2003) the intervention.

Multivariate analysis showed that prior exposure to third-generation cephalosporins (TGCs) (OR, 6.0; CI95, 1.9 to 18.6; P= .002) was an independent risk factor of ESBL-producing K. pneumoniae acquisition. The neurosurgical team was notified of the result, and the infectious diseases specialist visited the NSICU three times a week to regulate TGC use during the intervention period. Patients admitted before the intervention were older than patients admitted after. The respiratory acquisition of ESBL-producing K. pneumoniae per 1,000 patient-days (13.5 [CI95, 8.9 to 18.1] vs 2.7 [CI95, 0.9 to 4.6]) and the antimicrobial use density of TGCs (38.2 ± 5.0 vs 17.3 ± 2.6; P < .001) decreased significantly after the intervention.

Prior exposure to TGCs was an independent risk factor for the respiratory acquisition of ESBL-producing K. pneumoniae, and less use of TGCs was associated with a decrease in acquisition.

To evaluate the usefulness of screening cultures in the control of extended-spectrum beta-lactamase (ESBL)-producing Enterobacteriaceae in intensive care units (ICUs).

A 4-year retrospective study.

Two adult ICUs of a university-affiliated public hospital in France.

A total of 7,777 specimens were analyzed and 28 (0.97%) of 2,883 screened patients had a positive result on a screening test, among the 3,678 admitted patients. Thirteen of these 28 patients were only carriers; 4 were carriers and then were colonized or infected 2, 2, 3, and 8 days later, respectively; and 11 were colonized or infected before a screening test was positive. Cluster analysis showed that the occurrence of ESBL-producing Enterobacteriaceae cross-transmission within both ICUs was limited to 9 cases. Thus, most cases (19 of 28) were probably imported. Surveillance cultures failed to detect 9 of the 19 cases.

The low prevalence of ESBL-producing Enterobacteriaceae carriers on admission (0.45%) and the relative ineffectiveness of our screening test to detect imported cases suggest that systematic detection of ESBL-producing Enterobacteriaceae in ICU patients is not cost-effective and that the use of clinical cultures may be sufficient to control ESBL-producing Enterobacteriaceae in non-epidemic situations.

To examine and quantify the temporal trends of nosocomial multidrug-resistant, gram-negative bacilli.

A 9-year surveillance study was conducted. Multidrug resistance was defined as resistance to 3 or more antimicrobial classes.

Tertiary-care institution.

From 1994 to 2002, multidrug-resistant, gram-negative bacilli increased from 1% to 16% for multidrug-resistant Pseudomonas aeruginosa, 4% to 13% for multidrug-resistant Enterobacter species, 0.5% to 17% for multidrug-resistant Klebsiella species, 0% to 9% for multidrug-resistant Proteus species, and 0.2% to 4% for multidrug-resistant Escherichia coli (P ≤ .05). The most common pattern of multidrug resistance was co-resistance to quinolones, third-generation cephalosporins, and aminoglycosides.

The rapid rise of multidrug-resistant, gram-negative bacilli may warrant infection control programs to include these pathogens in strategies aimed at limiting the emergence and spread of antimicrobial-resistant pathogens.

To evaluate three different DNA techniques for typing nonfermentative gram-negative bacilli isolated from Latin American hospitals.

One hundred twenty-six nonfermentative gram-negative bacilli were typed.

Pseudomonas aeruginosa (n = 64) and Acinetobacter baumannii (n = 42) samples were obtained from blood cultures of patients admitted to 10 medical centers in Latin America during 1998 and Stenotrophomonas maltophilia (n = 20) samples were obtained from patients admitted to the Hospital São Paulo between 1999 and 2001.

All samples were typed using automated ribotyping, PFGE, and ERIC-PCR. The discriminatory power for each technique was calculated using Hunter's generalized formula.

All strains could be typed by automated ribotyping and ERIC-PCR, but two strains (1.6%) were not typeable by PFGE. All three techniques showed 100% reproducibility. The time to obtain the results was shorter for automated ribotyping and ERIC-PCR compared with PFGE. Likewise, the costs for ERIC-PCR and PFGE were lower than those for automated ribotyping. The interpretation of results was more complicated and more difficult with ERIC-PCR than with both PFGE and automated ribotyping. All techniques presented excellent discriminatory power for P. aeruginosa (0.98). PFGE presented the highest discriminatory power (0.94) for A. baumannii, and both PFGE and ERIC-PCR showed higher discriminatory power (0.90 for both) than automated ribotyping (0.82) for S. maltophilia.

PFGE showed the highest discriminatory power for typing these nonfermentative gram-negative bacilli. However, automated ribotyping and ERIC-PCR can provide results in a shorter time period with similar discriminatory power.

To investigate environmental contamination by extended-spectrum beta-lactamase-producing Enterobacteriaceae and to perform a comparative molecular analysis of clinical and environmental strains.

A 17-bed cardiac surgery intensive care unit of a 480-bed university teaching hospital.

Following an outbreak of extended-spectrum beta-lactamase-producing Enterobacteriaceae, an environmental survey revealed extensive contamination of the environment (particularly faucets, sink drains, and the joints of the countertops) by extended-spectrum beta-lactamase-producing Enterobacteriaceae. Environmental strains were compared with clinical strains by pulsed-field gel electrophoresis and randomly amplified polymorphic DNA.

A total of 62 environmental strains belonging to 4 species of extended-spectrum beta-lactamase-producing Enterobacteriaceae were analyzed and compared with 43 clinical strains obtained from 34 patients. Comparative molecular analysis revealed 4 identical or closely related patterns (3 from Klebsiella oxytoca and 1 from Enterobacter cloacae) between environmental and clinical strains.

Moist surfaces may serve as sources of multiply resistant Enterobacteriaceae in the intensive care unit. Identification and disinfection of such sources may therefore be helpful in prevention and control of outbreaks.

Nosocomial infections due to medical devices are of increasing concern to infection control practitioners. Attempts to prevent such infections have included surveillance cultures of endoscopes and bronchoscopes. In July 2002, the infectious disease consultation service was asked to see three patients with sepsis due to multidrug-resistant Pseudomonas aeruginosa after endoscopic retrograde cholangiopancreatography (ERCP).

To describe an outbreak of multidrug-resistant P. aeruginosa sepsis after ERCP at an institution that performs routine surveillance cultures of endoscopes.

A traditional outbreak investigation supplemented by pulsed-field gel electrophoresis (PFGE) was undertaken, including a case-control analysis based on the hypothesis that all infected individuals had their ERCP performed with the same endoscope.

A tertiary-care academic medical center.

The case-control analysis confirmed the hypothesis that undergoing ERCP with the implicated endoscope was associated with a culture positive for Pseudomonas (P = .01). The available strains were identical by PFGE. This outbreak occurred despite a negative surveillance culture of the implicated endoscope 1 month earlier.

Infectious morbidity can occur after endoscopy despite negative surveillance cultures. The practice of routine endoscope cultures does not prevent device-related infectious morbidity.

To evaluate risk factors and treatment outcomes of bloodstream infections caused by extended-spectrum beta-lactamase-producing Klebsiella pneumoniae (ESBL-KP).

Retrospective case-control study. Stored blood isolates of K. pneumoniae were tested for ESBL production by NCCLS guidelines, double-disk synergy test, or both.

A 1,500-bed, tertiary-care university hospital and referral center.

Sixty case-patients with bacteremia due to ESBL-KP were compared with 60 matched control-patients with non-ESBL-KP.

There were no significant differences in age, gender, APACHE II score, or underlying diseases between the groups. Independent risk factors for infections caused by ESBL-KP were urinary catheterization, invasive procedure within the previous 72 hours, and an increasing number of antibiotics administered within the previous 30 days. Complete response rate, evaluated 72 hours after initial antimicrobial therapy, was higher among control-patients (13.3% vs 36.7%; P = .003). Treatment failure rate was higher among case-patients (35.0% vs 15%; P = .011). Overall 30-day mortality rate was 30% for case-patients and 28.3% for control-patients (P = .841). Case-patients who received imipenem or ciprofloxacin as a definitive antibiotic had 10.5% mortality. The mortality rate for initially ineffective therapy was no higher than that for initially effective therapy (9.1% vs 11.1%; P = 1.000), but statistical power was low for evaluating mortality in the absence of septic shock.

For K. pneumoniae bacteremia, patients with ESBL-KP had a higher initial treatment failure rate but did not have higher mortality if antimicrobial therapy was appropriately adjusted in this study with limited statistical power.

To evaluate the emergence of resistance of Pseudomonas aeruginosa and Acinetobacter species to imipenem, ciprofloxacin, or both after the use of these drugs and to compare resistant with susceptible isolates by molecular typing.

Cohort study.

Burn intensive care unit (ICU) with 4 beds in a tertiary-care university hospital.

During 16 months, surveillance cultures were performed for all patients admitted to the ICU. Demographic information was obtained for each patient. Molecular typing was done by pulsed-field gel electrophoresis using restriction enzymes for 71 isolates of P. aeruginosa and Acinetobacter species.

Thirty-four patients were admitted and 22 were colonized by susceptible P. aeruginosa or Acinetobacter species before they used the antimicrobials. Nine (41%) of these patients had a resistant isolate after antimicrobial use: 5 had used imipenem alone, 1 had used ciprofloxacin, and 3 had used both drugs. The interval between isolation of the susceptible and resistant isolates ranged from 4 to 25 days, but was 10 or more days for 6 patients. Molecular typing revealed that susceptible and resistant isolates from each patient were different and that although there were no predominant clones among susceptible isolates, there was a predominant clone among resistant isolates of P. aeruginosa and of Acinetobacter.

Resistance was not due to the acquisition of resistance mechanisms by a previously susceptible strain, but rather to cross-transmission. Although various measures involving antimicrobial use have received great attention, it would seem that practices to prevent cross-transmission are more important in controlling resistance.

To determine the incidence of urinary tract infections (UTIs) following transrectal ultrasound guided needle biopsy of the prostate (TRUBP) and the bacteriology of these infections.

Retrospective evaluation of the charts and records of all patients who underwent TRUBP between June 1, 2002, and August 31, 2003.

American University of Beirut Medical Center, a tertiary-care center in Lebanon.

Two hundred seven patients underwent TRUBP. All received prophylactic antibiotics. One hundred twenty (58%) received ciprofloxacin alone, whereas 87 (42%) received both ciprofloxacin and gentamicin. Sixty-one patients (29.5%) had an enema prior to the procedure, whereas 146 (70.5%) did not.

Thirteen patients (6.3%) were admitted with UTI. All had rigors and fever on admission. Symptoms appeared at a mean of 2.7 days and the mean hospital stay was 9.2 days. The mean duration of antibiotic treatment was 23.2 days. Ten (77%) of the patients had positive bacteriology. Urine cultures were positive in 8 (61.5%) of the patients and blood cultures in 6 (46.2%). All positive cultures grew Escherichia coli resistant to ciprofloxacin, with 5 isolates producing extended-spectrum beta-lactamases.

TRUBP continues to be associated with significant infectious complications, especially UTI. Given the increasing incidence of antibiotic resistance mainly among the Enterobacteriaceae, antimicrobial prophylaxis practices should be reevaluated and the universal administration of quinolones alone or in combination with aminoglycosides should be reconsidered.

On February 19, 2003, four patients (patients 1-4) in the neurology ward underwent cranial magnetic resonance angiography (MRA) and developed fever within 1 hour afterward. Klebsiella oxytoca was isolated from blood cultures of patients 1 through 3.

To identify the source of this cluster of nosocomial K. oxytoca bloodstream infections.

Outbreak investigation.

A 1,000-bed university hospital.

The infection control team reviewed patient charts and interviewed nursing staff about the preparation and administration of parenteral fluids. The procedure of cranial MRA was observed. Arbitrarily primed polymerase chain reaction (AP-PCR) was performed to show the clonal relationship among these three strains.

AP-PCR revealed that three K. oxytoca isolates had the same molecular profile. Cranial MRA was found to be the only common source among these patients. During MRA, before injection of the contrast medium, normal saline solution was infused to check the functioning of the intravenous catheter. Use of the solution for multiple patients was routine, but the access diaphragm of the bottle was not cleansed. The bottle of normal saline solution used on February 19 had already been discarded and the culture sample taken from the solution on the day of observation was sterile.

We speculate that normal saline solution became contaminated during manipulation and that successive uses might have been responsible for this cluster. Poor aseptic techniques employed during successive uses appear to be the most likely route of contamination. Use of parenteral solutions for multiple patients was discontinued..