To demonstrate the effectiveness of the Charlson weighted index of comorbidity (WIC) for controlling comorbidity in prospective studies focusing on mortality in patients with Staphylococcus aureus bacteremia (SAB).

Cohort study.

Two tertiary-care, university-affiliated hospitals in France.

One hundred sixty-six inpatients 18 years or older consecutively diagnosed with SAB from May 15, 2001, to May 15, 2002.

Patients were prospectively assessed and cases were followed by the infectious diseases consult service at least 3 months after effective antibiotic therapy completion. The Charlson WIC was computed and dichotomized into scores of fewer than 3 points and 3 or more points. Bacteremia source, acute complication due to SAB acquisition in the ICU, and inappropriate empiric antibiotic therapy were recorded. The endpoint was death due to SAB and overall mortality.

In univariate analysis, the Charlson WIC was able to predict overall mortality and S. aureus-related death. The following variables were found to be independently predictive of mortality due to SAB using the Cox model: an acute complication due to S. aureus (OR, 8.9; CI95, 4 to 19.7; P < .001), a Charlson WIC score of 3 or more (OR, 3; CI95, 1.3 to 5.5; P = .006), and age (OR, 1.04; CI95, 1.009 to 1.07; P < .01).

Comorbidity contributes to death in patients with SAB. The Charlson WIC is a good predictor of mortality in this population and may be a useful instrument to control comorbidity in studies aiming to investigate risk factors for death due to bacteremia.

To evaluate the outcome of attempted Hickman catheter salvage in neutropenic cancer patients with Staphylococcus aureus bacteremia who were not using antibiotic lock therapy.

Retrospective cohort study.

A university-affiliated, tertiary-care hospital with 1,500 beds for adult patients.

All neutropenic cancer patients who had a Hickman catheter andS. aureus bacteremia (32 episodes in 29 patients) between January 1998 and March 2002.

Salvage attempts were defined as cases where the Hickman catheter was not removed until we obtained the results of follow-up blood cultures performed 48 to 72 hours after starting treatment with antistaphylococcal antibiotics. Salvage was considered to be successful if the Hickman catheter was still in place 3 months later without recurrent bacteremia or death.

Catheter salvage was attempted in 24 (75%) of the 32 episodes. Of the salvage attempts, the success rate was 50% (12 of 24). Salvage attempts were successful in 14% (1 of 7) of the episodes with positive follow-up blood cultures, and in 65% (11 of 17) of those with negative follow-up blood cultures (P = .07). If the analysis is confined to cases with no external signs of catheter infection, salvage attempts were successful in 14% (1 of 7) of the episodes with positive follow-up blood cultures and in 80% (8 of 10) of those with negative follow-up blood cultures (P = .02).

In neutropenic cancer patients with S. aureus bacteremia, attempted catheter salvage without antibiotic lock therapy was successful in 50% of the cases.

To examine the clinical and epidemiologic features, excess length of stay, extra costs, and mortality attributable to bloodstream infection (BSI) in neutropenic patients with hematologic malignancies.

Prospective cohort and matched case-control study.

All adult neutropenic patients with hematologic malignancies admitted to Cologne University Hospital between May 1, 1997, and April 30, 1998, were prospectively observed. Case-patients were defined as patients with nosocomial BSI; control-patients were selected among patients without BSI.

During the study period, the BSI rate in neutropenic patients was 14.3 per 100 neutropenic episodes. Eighty-four case-patients were included. Matching was successful for 96% of the cohort; 81 matched pairs were studied. The mean total length of stay was significantly longer for patients with BSI than for control-patients (37 vs 29 days; P = .002). Extra costs attributable to the infection averaged $3,200 (U.S.) per patient. The crude mortality rates of case-patients and control-patients were 16% and 4%, respectively (P = .013), with an attributable mortality of 12% (odds ratio, 11). Eighty-seven percent of patients met the criteria for sepsis according to the American College of Chest Physicians/Society of Critical Care Medicine. Severe sepsis or septic shock occurred in 13% of patients and was correlated with mortality (55% vs 10% in patients without severe sepsis or septic shock; P =.01).

Nosocomial BSI in neutropenic patients is significantly associated with an excess length of hospital stay, extra costs, and excess mortality. Severe sepsis and septic shock are closely correlated with an adverse outcome.

To evaluate excess mortality in critically ill patients with Escherichia coli bacteremia after adjustment for severity of illness.

Retrospective (1992-2000), pairwise-matched (1:2), risk-adjusted cohort study.

Fifty-four-bed ICU in a university hospital including a medical and surgical ICU, a unit for care after cardiac surgery, and a burns unit.

ICU patients with nosocomial E. coli bacteremia (defined as cases; n = 64) and control-patients without nosocomial bloodstream infection (n = 128).

Case-patients were matched with control-patients on the basis of the Acute Physiology and Chronic Health Evaluation (APACHE) II system: an equal APACHE II score (± 2 points) and diagnostic category. In addition, control-patients were required to have an ICU stay at least as long as that of the respective case-patients prior to onset of the bacteremia.

The overall rate of appropriate antibiotic therapy in patients with E. coli bacteremia was high (93%) and such therapy was initiated soon after onset of the bacteremia (0.6 ± 1.0 day). ICU patients with E. coli bacteremia had more acute renal failure. No differences were noted between case-patients and control-patients in incidence of acute respiratory failure, hemodynamic instability, or age. No differences were observed in length of mechanical ventilation or length of ICU stay. In-hospital mortality rates for cases and controls were not different (43.8% and 45.3%, respectively; P = .959).

After adjustment for disease severity and acute illness and in the presence of adequate antibiotic therapy, no excess mortality was found in ICU patients with E. coli bacteremia.

Defining risk factors for central venous catheter (CVC)-associated bloodstream infections (BSIs) is critical to establishing prevention measures, especially for factors such as nurse staffing and antimicrobial-impregnated CVCs.

We prospectively monitored CVCs, nurse staffing, and patient-related variables for CVC-associated BSIs among adults admitted to eight ICUs during 2 years.

A total of 240 CVC-associated BSIs (2.8%) were identified among 4,535 patients, representing 8,593 CVCs. Antimicrobial-impregnated CVCs reduced the risk for CVC-associated BSI only among patients whose CVC was used to administer total parenteral nutrition (TPN, 2.6 CVC-associated BSIs per 1,000 CVC-days vs no TPN, 7.5 CVC-associated BSIs per 1,000 CVC-days; P = .006). Among patients not receiving TPN, there was an increase in the risk of CVC-associated BSI in patients cared for by “float” nurses for more than 60% of the duration of the CVC. In multivariable analysis, risk factors for CVC-associated BSIs were the use of TPN in non-antimicrobial-impregnated CVCs (P = .0001), patient cared for by a float nurse for more than 60% of CVC-days (P = .0019), no antibiotics administered to the patient within 48 hours of insertion (P = .0001), and patient unarousable for 70% or more of the duration of the CVC (P = .0001). Peripherally inserted central catheters (PICCs) were associated with a lower risk for CVC-associated BSI (P = .0001).

Antimicrobial-impregnated CVCs reduced the risk of CVC-associated BSI by 66% in patients receiving TPN. Limiting the use of float nurses for ICU patients with CVCs and the use of PICCs may also reduce the risk of CVC-associated BSI.

To describe the conceptual framework and methodology of the Evaluation of Processes and Indicators in Infection Control (EPIC) study and present results of CVC insertion characteristics and organizational practices for preventing BSIs. The goal of the EPIC study was to evaluate relationships among processes of care, organizational characteristics, and the outcome of BSI.

This was a multicenter prospective observational study of variation in hospital practices related to preventing CVC-associated BSIs. Process of care information (eg, barrier use during insertions and experience of the inserting practitioner) was collected for a random sample of approximately 5 CVC insertions per month per hospital during November 1998 to December 1999. Organization demographic and practice information (eg, surveillance activities and staff and ICU nurse staffing levels) was also collected.

Medical, surgical, or medical-surgical ICUs from 55 hospitals (41 U.S. and 14 international sites).

Process information was obtained for 3,320 CVC insertions with an average of 58.2 (± 16.1) insertions per hospital. Fifty-four hospitals provided policy and practice information.

Staff spent an average of 13 hours per week in study ICU surveillance. Most patients received nontunneled, multiple lumen CVCs, of which fewer than 25% were coated with antimicrobial material. Regarding barriers, most clinicians wore masks (81.5%) and gowns (76.8%); 58.1% used large drapes. Few hospitals (18.1%) used an intravenous team to manage ICU CVCs.

Substantial variation exists in CVC insertion practice and BSI prevention activities. Understanding which practices have the greatest impact on BSI rates can help hospitals better target improvement interventions.

To examine the current status of bloodstream infections (BSIs) in a community hospital as part of a 25-year longitudinal study.

Retrospective descriptive epidemiologic study.

Community teaching hospital.

All inpatients in 1998 with a positive blood culture who met the CDC NNIS System case definition of BSI.

Cases were stratified by underlying illness category using case mix adjustment categories (after McCabe) and reviewed for associations among mortality, underlying illness severity, and multiple clinical and laboratory parameters.

Of 19,289 patients discharged in 1998,185 had an episode of infection documented by blood culture (96 cases per 10,000 inpatients). BSI was twice as frequent in patients 65 years and older compared with younger patients. BSIs caused or contributed to the deaths of 22 patients for an overall case-fatality rate of 11.9% compared with 20.7% in 1982 (P = .02). Striking decreases were noted for in-hospital patient mortality in 1998 for BSIs with ultimately and rapidly fatal underlying illnesses (P = .02 and P < .10, respectively). Primary bacteremia decreased compared with 1982. Antibiotic use was vigorous, but resistance was modest in both nosocomial and community-acquired organisms and had changed little from 1982 and 1987.

Compared with previous studies, case-fatality rates in patients with BSI were substantially lower in rapidly fatal and ultimately fatal underlying illness categories. Antibiotic use was extensive but prompt and appropriate. Microorganism resistance to antibiotics changed little from the 1980s.

To determine the prevalence of central venous catheter (CVC) use among patients both within and outside the ICU setting.

A 1-day prevalence survey of CVC use among adult inpatients at six medical centers participating in the Prevention Epicenter Program of the CDC. Using a standardized form, observers at each Epicenter performed a hospital-wide survey, collecting data on CVC use.

Inpatient wards and ICUs of six large urban teaching hospitals.

At the six medical centers, 2,459 patients were surveyed; 29% had CVCs. Among the hospitals, from 43% to 80% (mean, 59.3%) of ICU patients and from 7% to 39% (mean, 23.7%) of non-ICU patients had CVCs. Despite the lower rate of CVC use on non-ICU wards, the actual number of CVCs outside the ICUs exceeded that of the ICUs. Most catheters were inserted in the subclavian (55%) or jugular (22%) site, with femoral (6%) and peripheral (15%) sites less commonly used. The jugular (33.0% vs 16.6%; P < .001) and femoral (13.8% vs 2.7%; P < .001) sites were more frequently used in ICU patients, whereas peripherally inserted (19.9% vs 5.9%; P < .001) and subclavian (60.7% vs 47.3%; P < .001) catheters were more commonly used in non-ICU patients.

Current surveillance and infection control efforts to reduce morbidity and mortality associated with bloodstream infections concentrate on the high-risk ICU patients with CVCs. Our survey demonstrated that two-thirds of identified CVCs were not in ICU patients and suggests that more efforts should be directed to patients with CVCs who are outside the ICU.

Efficient methods are needed to monitor infections associated with long-term central venous catheters (CVCs) in both inpatient and outpatient settings. Automated medical records and claims data have been used for surveillance of these infections without evaluation of their accuracy or validity.

To determine the feasibility of using electronic records to identify CVC placement and design a system for identifying CVC-associated infections.

Retrospective cohort study at an HMO and two teaching hospitals in Boston, one adult (hospital A) and one pediatric (hospital B), between January 1991 and December 1997. Tunneled catheters, totally implanted catheters, and hemodialysis catheters were examined. Claims databases of both the HMO and the hospitals were searched for 10 CPT codes, 2 ICD-9 codes, and internal charge codes indicating CVC insertion. Lists were compared with each other and with medical records for correlation and accuracy.

All members of the HMO who had a CVC inserted at one of the two hospitals during the study period.

There was wide variation in the CVC insertions identified in each database. Although ICD-9 codes at each hospital and CPT/ICD-9 combinations at the HMO found similar total numbers of CVCs, there was little overlap between the individuals identified (62% for hospital A with HMO and 4% for hospital B).

Claims data from different sources do not identify the same CVC insertion procedures. Current administrative databases are not ready to be used for electronic surveillance of CVC-associated complications without extensive modification and validation.

To evaluate whether a natural language processing system, SymText, was comparable to human interpretation of chest radiograph reports for identifying the mention of a central venous catheter (CVC), and whether use of SymText could detect patients who had a CVC.

To identify patients who had a CVC, we performed two surveys of hospitalized patients. Then, we obtained available reports from 104 patients who had a CVC during one of two cross-sectional surveys (ie, case-patients) and 104 randomly selected patients who did not have a CVC (ie, control-patients).

A 600-bed public teaching hospital.

Chest radiograph reports were available from 124 of the 208 participants. Compared with human interpretation, SymText had a sensitivity of 95.8% and a specificity of 98.7%. The use of SymText to identify case- and control-patients resulted in a sensitivity of 43% and a specificity of 98%. Successful application of SymText varied significantly by venous insertion site (eg, a sensitivity of 78% for subclavian and a sensitivity of 3.7% for femoral). Twenty-six percent of the case-patients had a femoral CVC.

Compared with human interpretation, SymText performed well in interpreting whether a report mentioned a CVC. In patient populations with less frequent CVC placement in femoral veins, the sensitivity for CVC detection likely would be higher. Applying a natural language processing system to chest radiograph reports may be a useful adjunct to other data sources to automate detection of patients who had a CVC.

To report the results of our preintervention investigation and subsequent 19-month three-phase intervention study designed to reduce pyrogenic reactions among patients undergoing cardiac catheterization using reprocessed catheters.

A case-control study for the preintervention period and a prospective cohort study for the intervention period.

A 400-bed hospital in Belo Horizonte, Brazil.

Any patient undergoing cardiac catheterization in the hospital.

Three intervention phases were implemented to improve the quality of the water supplied to the cardiac catheter reprocessing laboratory. Standard operating procedures for reprocessing cardiac catheters were established and reprocessing staff were trained and educated.

The rate of pyrogenic reactions decreased significantly during the intervention phases, from 12.8% (159 of 1,239) in phase 1 to 5.3% (38 of 712) in phase 2 to 0.5% (4 of 769) in phase 3 (chi-square test for linear trend, 97.5; P < .001).

Improving water quality and using standard operating procedures for reprocessing catheters can prevent pyrogenic reactions in hospitalized patients.