To assess changes in the epidemiology of nosocomial candidemia in the post-fluconazole era among hospitalized patients using a case–control study design.

Candidemia case-patients were matched 1:1 on diagnosis, age, and length of stay with control-patients. Conditional logistic regression was used to determine predictors and outcomes of candidemia. Treatment regimens and compliance with national practice guidelines were compared among case-patients.

Barnes-Jewish Hospital, a 1,278-bed, tertiary-care center affiliated with Washington University School of Medicine, St. Louis, Missouri.

Patients admitted from January 1 to December 31, 2000. Case-patients were identified through the hospital microbiological surveillance system and matched with control-patients.

Predictors of candidemia included Hickman catheters (odds ratio [OR], 9.53; 95% confidence interval [CI95], 1.34 to 68.01), gastric acid suppressants (OR, 6.38; CI95, 2.33 to 17.43), nasogastric tubes (OR, 3.69; CI95, 1.27 to 10.78), antibiotics (OR, 1.46; CI95,1.15 to 1.86), and admission to the intensive care unit (OR, 6.40; CI95, 2.12 to 19.31). The crude case-fatality rate was 40%. Seventeen (15%) of the case-patients received the recommended treatment regimen according to recently published practice guidelines.

The epidemiology of candidemia has changed little at our hospital during the past decade and remains a significant cause of mortality. Further studies on the benefits of preventive therapy will be essential to improve the outcome of this infection.

To compare the epidemiology and genetic relatedness of Candida tropicalis isolates causing bloodstream infection (BSI) in two hospitals.

Two tertiary-care hospitals in Korea.

A retrospective molecular epidemiologic analysis using pulsed-field gel electrophoresis (PFGE) was performed with 49 C. tropicalis isolates from sporadic cases of BSI. The isolates were collected from 27 patients at Chonnam National University Hospital (CUH) during a 6-year period and 22 patients at Asan Medical Center (AMC) during a 2-year period.

Based on the PFGE patterns, the average similarity value (SAB) for the 27 isolates from CUH was 0.84 ± 0.08, which was significantly higher than that for the 22 isolates from AMC (0.78 ± 0.06; P < .001). Of the 49 strains from patients at the 2 hospitals, 9 isolates were placed into 3 subtypes with SAB values of 1.0, which indicated that they were identical. All 9 of these strains were isolated from CUH patients, and each type strain was isolated sporadically during a period ranging from 4 months to 3 years. On comparison of the clinical characteristics of the patients of the 2 hospitals, the CUH strains were isolated more frequently from non-neutropenic patients and patients with central venous catheter–related fungemia; cases from CUH had a better outcome than those from AMC (P < .05).

These data show that the clinical and epidemiologic characteristics of C. tropicalis fungemia may differ markedly among hospitals and that some cases of C. tropicalis fungemia may be caused by endemic strains within a hospital.

To investigate an outbreak of Candida parapsilosis bloodstream infections (BSIs) involving three patients admitted to a pediatric oncology unit between April and June 2002.

After the third case was documented, cultures were performed of the hands of all medical and paramedical staff members in the pediatric oncology unit and of environmental surfaces in the rooms occupied by the three patients. Electrophoretic karyotyping with pulsed-field gel electrophoresis and arbitrarily primed polymerase chain reaction were used to assess the genetic relatedness among C. parapsilosis isolates.

The three cases of C. parapsilosis BSI were diagnosed based on blood cultures performed during a 38-day period. Evidence of prior C. parapsilosis colonization of the gastrointestinal tract was present in only the first case. Each patient had an indwelling central venous catheter (CVC), which was promptly removed, and semiquantitative catheter tip cultures also revealed C. parapsilosis. None of the 30 environmental cultures performed was positive for C. parapsilosis, but the fungus was isolated from the hands of 6 of the 20 nurses tested. Both molecular typing methods revealed identical DNA fingerprinting patterns for all 13 patient isolates (7 from blood, 3 from CVC tips, and 3 from the gastrointestinal tract) and for 5 of the 6 recovered from the nurses' hands.

These findings suggest the possibility of cross-infection with a single C. parapsilosis strain that was transmitted (probably during CVC dressing changes) by nurses whose hands were colonized with it. The role of previous gastrointestinal colonization in the first case cannot be excluded.

To study the characteristics of catheter-related, gram-negative bacteremia (GNB) and the role of central venous catheter (CVC) removal.

This retrospective study involved a search of the microbiological department records of CVC and blood cultures and patients' medical records.

University of Texas M. D. Anderson Cancer Center, a tertiary-care hospital in Houston, Texas.

Patients with cancer who had catheter-related GNB, defined as (1) a positive catheter tip culture with at least 15 colony-forming units semiquantitatively, (2) isolation of the same organism from the tip and peripheral blood cultures, (3) no other source for bacteremia except the CVC, and (4) clinical manifestations of infection (fever or chills).

Between January 1990 and December 1996, 72 cases of catheter-related GNB were available for review. Most of the patients (67; 93%) had their CVCs removed in response to the bacteremia. Few patients (5; 7%) retained their CVCs and were treated with appropriate antibiotics. When CVCs were removed, only 1 patient (1%) relapsed with the same organism, whereas all 5 patients with retained CVCs relapsed after having responded (P < .001). The most commonly isolated organisms were Enterobacter, Klebsiella, Stenotrophomonas, Pseudomonas, and Acinetobacter species. Catheter removal within 72 hours of the onset of the catheter-related GNB was the only independent protective factor against relapse of the infection (odds ratio, 0.13; 95% confidence interval, 0.02–0.75; P = .02).

In patients with documented catheter-related GNB, CVCs should be removed within 48 to 72 hours to prevent relapse.

To determine the effect of routine intravenous (IV) administration set changes on central venous catheter (CVC) colonization and catheter-related bacteremia.

Prospective, randomized, controlled trial.

Eighteen-bed intensive care unit (ICU) in a large metropolitan hospital.

Two hundred fifty-one patients with 404 chlorhexidine gluconate and silver sulfadiazine–coated multilumen CVCs.

CVCs inserted in the ICU and in situ on day 4 were randomized to have their IV administration sets changed on day 4 (n = 203) or not at all (n = 201). Use of fluid containers and blood product administration sets was limited to 24 hours. CVCs were removed when not required, infection was suspected, or in place on day 7. Catheter cultures were performed on removal by blinded laboratory staff. Catheter-related bacteremia was diagnosed by a blinded intensivist using strict definitions. Data were collected regarding catheter duration, site, Acute Physiology and Chronic Health Evaluation (APACHE) II score, patient age, diagnosis, hyperglycemia, hypoalbuminemia, immune status, number of fluid containers and IV injections, and administration of propofol, blood, total parenteral nutrition, or lipid infusion.

There were 10 colonized CVCs in the group receiving a set change and 19 in the group not receiving one. This difference was not statistically significant on Kaplan–Meier survival analysis. There were 3 cases of catheter-related bacteremia per group. Logistic regression found that burns diagnosis and increased ICU stay significantly predicted colonization.

IV administration sets can be used for 7 days in patients with short-term, antiseptic-coated CVCs.

To identify the independent predictors of nosocomial bloodstream infections (BSIs) among critically ill adult trauma patients.

A prospective, cohort design was used to study patients who met predetermined inclusion criteria. Basic descriptive and univariate statistical analyses were performed to identify unadjusted predictors. A forward stepwise multivariate logistic regression analysis was then conducted to identify independent predictors of nosocomial BSI.

Level I university-affiliated shock trauma center.

Three hundred sixty-one critically ill adult trauma patients, 55 of whom developed nosocomial BSIs (15.2%).

Data analysis of 45 variables indicated that only 9 were independent predictors of nosocomial BSI: presence of a chest tube, use of immunosuppressive agents, presence of microbial resistance, length of stay, presence of preexisting infection, percentage change of serum albumin levels, patient disposition, transfusion of 10 or more units of blood, and number of central venous catheters (CVCs) for patients who had 4 or more. The classification index of the final regression model at a cut-off point of 0.5 had a specificity of 97.4%, a sensitivity of 60%, a positive predictive value of 76.7%, a negative predictive value of 93%, and an overall precision of 91%.

In this study, only 9 variables were independent predictors of nosocomial BSI. Our findings are specific to critically ill adult trauma patients and should be interpreted within the context of this particular population.

To evaluate an antimicrobial dressing consisting of hydrophilic polyurethane foam with chlorhexidine gluconate for activity against several antibiotic-resistant clinical isolates as well as American Type Culture Collection reference strains using zone of inhibition assays.

Sterile foam samples with chlorhexidine gluconate and untreated controls were transferred onto inoculated agar plates. Plates were incubated at 35°C to 37° C for 24 hours and examined for zones of inhibition around the foam samples.

Polyurethane foam with chlorhexidine gluconate showed antimicrobial activity in vitro against all of the challenge organisms including antibiotic-resistant clinical isolates.

The data from this in vitro study support the hypothesis that polyurethane foam with chlorhexidine gluconate has an antimicrobial effect against antibiotic-resistant Staphylococcus and Enterococcus species, as well as Candida species.

To compare the costs with the benefits of using chlorhexidine gluconate dressings on central venous catheters and to determine the effectiveness of these dressings in reducing local infections and catheter-related bloodstream infections (CRBSIs), costs, and mortality.

Cost–benefit analysis using randomized, controlled trial data on chlorhexidine dressing prevention of local infection and CRBSI, data on cost of chlorhexidine dressing versus standard treatment, data on averted cost of treating local infection and CRBSI, and data on mortality attributable to CRBSI. Decision analysis evaluated averted CRBSI treatment cost per patient resulting from chlorhexidine dressing use. Sensitivity analyses demonstrated net benefit of chlorhexidine dressing, varying baseline rate of CRBSI, incremental cost of treating CRBSI, and number of catheters, and evaluated mortality preventable through chlorhexidine dressing use, varying baseline rate of CRBSI, number of catheters, and mortality attributable to CRBSI.

Patients of all Philadelphia area hospitals and one Philadelphia academic medical center.

Estimated potential annual U.S. net benefits from chlorhexidine dressing use ranged from $275 million to approximately $1.97 billion. Cost–benefit findings persisted in sensitivity analyses varying baseline rate of CRBSI, incremental cost of treating CRBSI, and overall number of catheters used. Preventable mortality analyses showed potential decreases of between 329 and 3,906 U.S. deaths annually as a result of nationwide use of chlorhexidine dressing.

Chlorhexidine dressings would reduce costs, local infections and CRBSIs, and deaths. Use of chlorhexidine dressings should be considered to prevent infections among patients with catheters.

To evaluate the incidence of nosocomial bacteremias related to the use of non-impregnated central venous catheters (CVCs) when only non-technologic strategies were used to prevent them.

This was a prospective study of infectious complications of CVCs placed in intensive care unit (ICU) patients from April 1997 to December 2001.

The medical–surgical ICU of a tertiary-care, university-affiliated hospital in Argentina.

We studied all patients admitted to the ICU using non-impregnated CVCs. Maximal sterile barrier precautions (ie, use of cap, mask, sterile gown, sterile gloves, and large sterile drape), strict handwashing, preparation of the patients' skin with antiseptic solutions, insertion and management of catheters by trained personnel, and continuing quality improvement programs aimed at appropriate insertion and maintenance of catheters were employed.

During the study period, 2,525 patients were admitted to the ICU. Eight hundred sixty-eight patients had 1,037 CVCs inserted. The number of CVC-related bloodstream infections (BSIs), acquired in the ICU, was 2.7 per 1,000 CVC-days (13 nosocomial CVC-related BSIs during 4,770 days of CVC use). Microorganisms isolated included methicillin-susceptible Staphylococcus aureus (n = 6), methicillin-resistant S. aureus (n = 2), coagulase-negative methicillin-resistant Staphylococcus (n = 2), Escherichia coli (n = 1), Klebsiella pneumoniae (n = 1), and Enterobacter cloacae (n = 1).

A low rate of catheter-related BSI was achieved without antimicrobial-impregnated catheters. The incidence of CVC-associated bacteremias corresponded to the 10th to 20th percentile range of the National Nosocomial Infections Surveillance System hospitals for the same type of ICU.

To determine risk factors for hemodialysis catheter-related bloodstream infections (HCRBSIs) and investigate whether use of maximal sterile barrier precautions would prevent HCRBSIs.

Tertiary-care medical center hemodialysis unit.

Open trial with historical comparison and case-control study of risk factors for HCRBSIs.

Prospective surveillance was used to compare HCRBSI rates for 1 year before and after implementation of maximal sterile barrier precautions. A case–control study compared 50 case-patients with HCRBSI with 51 randomly selected control-patients.

The HCRBSI rate was 1.6% per 100 dialysis runs (CI95, 1.1%–2.3%) in the first year and 0.77% (CI95, 0.5%–1.1%) in the second year (P = .0106). The most frequent cause of HCRBSI was MRSA in the first year (15 of 32) and MSSA in the second year (13 of 18). Ten MRSA blood isolates in the first year were identical by PFGE. Diabetes mellitus was a risk factor for HCRBSI. Age, gender, site of hemodialysis central venous catheter (CVC), other underlying diseases, coma score, APACHE II score, serum albumin level, and cholesterol level were not associated with HCRBSI and did not change between the 2 years. Hospital stay was prolonged for case-patients (32.78 ± 20.96 days) versus control-patients (22.75 ± 17.33 days), but mortality did not differ.

Use of maximal sterile barrier precautions during the insertion of CVCs reduced HCRBSIs in dialysis patients and seemed cost-effective. Diabetes mellitus was associated with HCRBSI. An outbreak of MRSA in the first year was likely caused by cross-infection via medical personnel.

To assess the influence of prophylactic selective bowel decontamination (SBD) on the spectrum of microbes causing bloodstream infection (BSI).

The microbes causing BSI in neutropenic patients of a hematologic ward (HW) and a bone marrow transplantation unit (BMTU), respectively, were compared by retrospective analysis of blood culture results from January 1996 to June 2003.

A 30-bed HW (no SBD) and a BMTU including a 7-bed normal care ward and an 8-bed intensive care unit (SBD used) of a 2,200-bed university teaching hospital.

The overall incidences of bacteremia in the HW and the BMTU were similar (72.6 vs 70.6 episodes per 1,000 admissions; P = .8). Two hundred twenty episodes of BSI were recorded in 164 neutropenic patients of the HW and 153 episodes in 127 neutropenic patients of the BMTU. Enterobacteriaceae (OR, 3.14; CI95, 1.67–5.97; P = .0002) and Streptococcus species (OR, 2.04; CI95, 1.14–3.70; P = .015) were observed more frequently in HW patients and coagulase-negative staphylococci more frequently in BMTU patients (OR, 0.15; CI95, 0.09–0.26; P< .00001). No statistically significant differences were found for gram-negative nonfermentative bacilli (P = .53), Staphylococcus aureus (P = .21), Enterococcus species (P = .48), anaerobic bacteria (P = .1), or fungi (P = .50).

SBD did not lead to a significant reduction in the incidence of bacteremia, but significant changes in microbes recovered from blood cultures were observed. SBD should be considered when empiric antimicrobial therapy is prescribed for suspected BSI.