National surveillance of nosocomial blood stream infection due to species of Candida other than Candida albicans: frequency of occurrence and antifungal susceptibility in the SCOPE Program. SCOPE Participant Group. Surveillance and Control of Pathogens of Epidemiologic

A national surveillance program of nosocomial blood stream infections (BSI) in the USA between April 1995 and June 1996 revealed that Candida was the fourth leading cause of nosocomial BSI, accounting for 8% of all infections. Forty-eight percent of 379 episodes of candidemia were due to species other than Candida albicans. The rank order of non-C. albicans species was C. glabrata (20%) > C. tropicalis (11%) > C. parapsilosis (8%) > C. krusei (5%) > other Candida spp. (4%). The species distribution varied according to geographic region, with non-C. albicans species predominating in the Northeast (54%) and Southeast (53%) regions, and C. albicans predominating in the Northwest (60%) and Southwest (70%) regions. In vitro susceptibility studies demonstrated that 95% of non-C. albicans isolates were susceptible to 5-fluorocytosine, and 84% and 75% were susceptible to fluconazole and itraconazole, respectively. Geographic variation in susceptibility to itraconazole, but not other agents, was observed. Isolates from the Northwest and Southeast regions were more frequently resistant to itraconazole (29-30%) than those from the Northeast and Southwest regions (17-18%). Molecular epidemiologic studies revealed possible nosocomial transmission (five medical centers). Continued surveillance for the presence of non-C. albicans species among hospitalized patients is recommended.

Measuring health-related quality of life among patients infected with human immunodeficiency virus

Optimizing health-related quality of life (HRQL) has become an important treatment focus for patients infected with human immunodeficiency virus (HIV). Consideration of HRQL is especially relevant with the development of new antiretroviral agents that have significant side effects. The measurement of HRQL is still in evolution, and several methods have been used to quantify HRQL in the HIV-infected population. A review of existing studies shows that HRQL scores do not always correlate with disease stage or health indices and that symptoms have a significant impact on HRQL. Studies have also revealed that certain therapies for HIV and opportunistic infections exact a significant cost in terms of HRQL. HRQL outcomes will play a major role in treatment decisions for HIV-infected patients and in the development and marketing of new pharmaceutical agents in the near future.

The epidemiology of hemorrhage related to cardiothoracic operations

OBJECTIVE

To define the epidemiology, risk factors, and unadjusted cost of hemorrhages related to cardiothoracic operations.

STUDY DESIGN

We conducted two case-control studies to evaluate the risk of hemorrhage following cardiothoracic operations. The definition of hemorrhage required one of the following: reoperation for bleeding, postoperative loss of greater than 800 mL of blood over 4 hours, or surgeon-diagnosed excessive intraoperative bleeding.

SETTING

The cardiothoracic surgery service of a university hospital.

RESULTS

Of 511 patients undergoing cardiothoracic operations, 93 (18%) met the definition of hemorrhage. In the first case-control study, 3 (14%) of 21 cases and 0 of 42 controls died (odds ratio [OR], 15.0; 95% confidence interval [CI95], 1.18-191.55). Compared with controls, cases received significantly more packed red blood cells intraoperatively (OR, 1.18/100 mL; CI95, 1.01-1.38), and significantly more platelets (OR, 3.26/100 mL; CI95, 1.47-7.26) and fresh frozen plasma (OR, 1.73/100 mL; CI95, 1.05-.84) in the intensive-care unit. Cases were more likely than controls to receive protamine postoperatively (OR, 3.74; CI95, 1.27-11.02). Previous sternotomy, preoperative aspirin or heparin, and preoperative laboratory values did not predict bleeding. The median unadjusted hospital cost was $3,458 higher for patients who suffered hemorrhage than for controls. To decrease costs, hetastarch (acquisition cost $45/500 mL) was substituted for albumin (acquisition cost $76/100 mL) in the pump priming solution (estimated possible cost savings, $7,000-$53,000/year). Because hemorrhage rates increased subsequently, we conducted a second case-control study that identified patient age (P=.02) and use of greater than 5 mL/kg of hetastarch (OR, 1.82) as risk factors for hemorrhage. The cost of treating hemorrhages exceeded all estimates of possible cost savings ($7,000-$53,000 per year).

CONCLUSIONS

Our definition of hemorrhage identified patients who required increased volumes of blood products and who had an increased crude mortality rate and a higher unadjusted cost of hospitalization. Patient age and hetastarch use were risk factors for hemorrhage. Efforts to save money by substituting less expensive products inadvertently may increase costs by increasing the probability of perioperative adverse events.

Staphylococcus aureus bacteremia: the cost-effectiveness of long-term therapy associated with infectious diseases consultation

OBJECTIVE

To investigate the cost-effectiveness of long-term therapy for Staphylococcus aureus bacteremia and to determine if an infectious diseases consultation affected the duration of therapy.

METHODS

A decision analysis was performed based on data from the literature. To determine if consultation was related to therapy duration, a retrospective cohort study was performed using tightly matched pairs.

RESULTS

The excess cost per life saved by long-term antibiotics was $500,000. The excess cost per life-year saved was $18,000. Nine pairs were matched. Patients who received consultation were more likely to receive long-term therapy than controls (median 41 days vs 15 days for controls, P = .04).

CONCLUSIONS

The estimated cost per life-year saved by long-term therapy was similar to other accepted medical interventions. Infectious diseases consultation can encourage prolonged duration of antibiotic therapy for S aureus bacteremia.

Nosocomial streptococcal blood stream infections in the SCOPE Program: species occurrence and antimicrobial resistance. The SCOPE Hospital Study Group

Nosocomial blood stream infections due to streptococci represent an increasingly important problem, particularly among neutropenic cancer patients. This problem is compounded by the emerging resistance to antimicrobial agents commonly used for empiric or prophylactic treatment of hospitalized patients. In this study, we examined the species distribution and antimicrobial susceptibility profile of 295 streptococcal nosocomial blood stream isolates from more than 30 U.S. medical centers (SCOPE National Surveillance Program). Streptococci accounted for 5.9% of all nosocomial blood stream isolates reported. The viridans group streptococci (VGS) were the most frequently isolated streptococci (50.8%), followed by the beta-haemolytic streptococci (31.9%) and pneumococci (13.2%). The beta-haemolytic streptococci were dominated by serogroup B strains (63%), followed by serogroups A and G. Of these organisms, 193 strains were referred for subsequent monitor susceptibility testing. Approximately 14% of S. pneumoniae, 9.2% of VGS, and 0% of beta-haemolytic streptococci were resistant to penicillin. Ceftriaxone was highly active against virtually all isolates (93-100% susceptible) except the VGS (77% susceptible). The rank order for activity of the four agents tested against the 193 isolates was vancomycin > ceftriaxone > penicillin > erythromycin. Importantly, 69% of the penicillin intermediate and resistant strains of VGS were also resistant to at least one additional antimicrobial (31% resistant to ceftriaxone, 51% resistant to erythromycin, 15% resistant to both ceftriaxone and erythromycin). The relatively poor activity of erythromycin against virtually all streptococci and the frequent association of macrolide resistance with penicillin resistance among the VGS suggests that both macrolides and beta-lactam agents might have limited value as prophylactic agents for dental procedures and in empiric or prophylactic use in neutropenic patients.

An outbreak of Candida parapsilosis prosthetic valve endocarditis

Candida parapsilosis, an important nosocomial pathogen and the most common species of Candida found on the hands of health care workers, is a rare cause of prosthetic valve endocarditis (PVE). From March through June 1994, four cases of C. parapsilosis PVE were diagnosed at a 400-bed community hospital. The mean time to presentation after valve replacement surgery was 148 days (range, 20 to 345). Three of the four patients died of complications of PVE. Multiple environmental cultures were performed, and only one was positive for C. parapsilosis. Cultures from the bypass pump, cell saver, cardioplegia solution, and subsequent valves were all negative. All valve replacements were performed by the same operating room team. Interviews with the surgeon and physician assistant, the only personnel involved in all cases, revealed that their hypoallergenic gloves were subject to frequent tears during valve replacement procedures, often requiring several glove changes per procedure. Hand cultures of personnel were obtained, and cultures from 20 individuals (26%) were positive for C. parapsilosis. Hand cultures of the surgeon and physician assistant obtained 8 months after the last case had surgery were negative for yeasts. Molecular typing of the 3 available case isolates, 14 epidemiologically unrelated patient isolates, 1 environmental isolate, and 20 hand isolates was performed by electrophoretic karyotyping and restriction endonuclease analysis of genomic DNA using restriction enzymes BssHII and EagI followed by pulsed field gel electrophoresis. The three case isolates were identical by restriction endonuclease analysis of genomic DNA, and two of the three shared the same electrophoretic karyotyping profile. The remaining patient, environmental, and hand isolates represented 29 different DNA types and were distinctly different from the case isolates. All of the isolates tested were susceptible to amphotericin B, 5FC, fluconazole, and itraconazole. The circumstantial evidence suggests the probability of glove tears during valve replacement surgery and subsequent transmission of C. parapsilosis to patients.

Nosocomial enterococcal blood stream infections in the SCOPE Program: antimicrobial resistance, species occurrence, molecular testing results, and laboratory testing accuracy. SCOPE Hospital Study Group

Characteristics of nosocomial enterococcal blood stream infection (NEBSI) isolates obtained from patients at 41 U.S. hospitals participating in the SCOPE Program were studied. Isolates from 480 episodes of NEBSI were characterized according to species and antimicrobial susceptibility profile. Selected isolates were also identified to species and vancomycin resistance genotype using polymerase chain reaction based methods. Polymerase chain reaction genotyping and ribotyping were used as genetic markers for molecular epidemiologic typing. Enterococci were the third most common cause of nosocomial blood stream infection in this study, accounting for 11.7% of all isolates reported. Enterococcus faecalis was the most common species (59.6%), followed by E. faecium (19.4%). Species identification errors involving E. faecium, E. durans, E. avium, and E. raffinosus were observed. Vancomycin resistance was observed in 36.4% of all participating medical centers and varied from 11.1% of medical centers in the Northwest to 60.9% of medical centers in the Southwest. Vancomycin-resistant enterococci accounted for 20.6% of NEBSI in the Northeast, 11.4% in the Southeast, 11.1% in the Southwest, and 9.5% in the Northwest regions. VanA genotypes predominated in the Northeast and Southwest, whereas vanA and vanB genotypes were equally prevalent in the Northwest and Southeast. Molecular typing studies identified strains that were unique to individual hospitals as well as strains that were prevalent in several different hospitals. NEBSI with vancomycin-resistant enterococci continues to escalate among hospitalized patients in all geographic areas of the USA.

Vancomycin-resistant Enterococcus raffinosus: molecular epidemiology, species identification error, and frequency of occurrence in a national resistance surveillance program

Enterococcal blood stream infections are the third most common among all nosocomial blood stream infections in the United States and the occurrence of glycopeptide (vancomycin, teicoplanin) resistance in these isolates has markedly increased. Control of hospital-acquired infections with vancomycin-resistant enterococci requires high quality antimicrobial susceptibility test methods and species identification procedures as a supplement to epidemiologic investigation and appropriate infection control procedures. In this report, bacteremias caused by Enterococcus avium (BioMerieux Vitek, Hazelwood, MO, USA) were observed to be Enterococcus raffinosus infections (six of eight cases; 1.1% of all cases) when reference biochemical identification methods were applied. The vancomycin-susceptible E. raffinosis (two strains) and E. avium (two strains) had unique phenotypic and genotypic molecular profiles. In contrast, four vancomycin-resistant E. raffinosus strains (van A by polymerase chain reaction) from a single institution had the same phenotypic and molecular (PCR, PFGE, ribotyping) pattern, indicating clonal dissemination among four patients over a 66-day period. Clinical laboratories should be aware of the high probability that van A genes may be transferred from Enterococcus faecium or Enterococcus faecalis to other more rarely encountered Enterococcus species. Also contemporary, widely used commercial identification systems may fail to accurately identify those rare species. Errors appear to be most prevalent for E. avium, Enterococcus durans, and E. raffinosus based on the experience of the SCOPE Program.

Inducible amp C beta-lactamase producing gram-negative bacilli from blood stream infections: frequency, antimicrobial susceptibility, and molecular epidemiology in a national surveillance program (SCOPE)

A surveillance study of nosocomial blood stream infections [Surveillance and Control of Pathogens of Epidemiologic Importance (SCOPE)] was conducted during a 14-month period in 1995 to 1996 in approximately 50 American medical centers. Among the 4725 blood stream infections, the etiologic agent was Enterobacter spp. in 230, Citrobacter freundii in 24, and Serratia marcescens in 65. The vast majority of these isolates (89%) had been sent to the University of Iowa including 198 Enterobacter spp. (46 Enterobacter aerogenes, 141 Enterobacter cloacae, 11 other Enterobacter spp.), 23 C. freundii, and 62 S. marcescens. Because these species are capable of producing Amp C beta-lactamase, we examined their susceptibility to 12 broad-spectrum antimicrobial agents. The frequency of resistance to ceftazidime and the molecular epidemiology of ceftazidime-resistant strains was also examined. Among the Enterobacter spp. and C. freundii isolates, resistance to third generation cephalosporins (ceftazidime, ceftriaxone) and broad-spectrum semisynthetic penicillins (piperacillin), with or without an enzyme inhibitor (piperacillin/tazobactam), was high, e.g., 35 to 50%. The S. marcescens isolates were quite susceptible to all agents tested. Both imipenem and cefepime were active against virtually all isolates tested including 84 stably derepressed Amp C-producing ceftazidime-resistant strains of Enterobacter spp. and C. freundii. The overall rank order of activity for the six best agents against these Amp C-producing strains was: imipenem (100% susceptible) > amikacin = cefepime (98.6%) > ciprofloxacin = gentamicin = ofloxacin (93.6 to 94.0%). Molecular typing studies of ceftazidime-resistant E. cloacae using an automated ribotyping system, as well as pulsed-field gel electrophoresis, indicated that although clonal spread of a single strain occurred in some of the medical centers, most of the episodes of bacteremia were caused by patient-unique strains. Control of these resistant organisms will require attention to microbiologic recognition of phenotypes, to infection control practices, and to limiting the overuse of certain extended spectrum beta-lactams.

Nosocomial Candida. Epidemiology, transmission, and prevention

The NNIS and the newly established SCOPE data indicate that the relative proportion of organisms causing nosocomial bloodstream infections has changed over the last decade, with Candida species now being firmly established as one of the most frequent agents. The epidemiology of nosocomial candidemia is continually being refined, but established predisposing factors including immunosuppression and malignancies, use of broad spectrum antibiotics, and use of indwelling central catheters remain as significant risk factors. The high cost of health care and greater attention to continuous quality improvement will stimulate better and more effective ways of diagnosing and treating candida infections using combined clinical and microbiologic acumen. There is room for optimism as newer antifungal agents with reduced toxicities have impact on therapy of candidal infections. Aggressive development of still more agents and reformulations of older agents continue in earnest. Even greater consolation comes from the increased awareness of lay and medical personnel alike regarding the appropriate and judicious use of antimicrobial agents.

Microbiological factors influencing the outcome of nosocomial bloodstream infections: a 6-year validated, population-based model

All patients (n = 1,745) with nosocomial bloodstream infection identified between 1986 and 1991 at a single 900-bed tertiary care hospital were studied to identify microbiological factors independently associated with mortality due to the infection. Patients were identified by prospective, case-based surveillance and positive blood cultures. Mortality rates were examined for secular trends. Prognostic factors were determined with use of univariate and multivariate analyses, and both derivation and validation sets were used. A total of 1,745 patients developed nosocomial bloodstream infection. The 28-day crude mortality was 22%, and crude in-hospital mortality was 35%. Factors independently (all P < .05) associated with increased 28-day mortality rates were older age, longer length of hospital stay before bloodstream infection, and a diagnosis of cancer or disease of the digestive system. After adjustment for major confounders, Candida species were the only organisms independently influencing the outcome of nosocomial bloodstream infection (odds ratio [OR] for mortality = 1.84; 95% confidence interval [CI], 1.22-2.76; P = .0035). The two additional microbiological factors independently associated with increased mortality were pneumonia as a source of secondary infection (OR = 2.74; 95% CI, 1.87-4.00; P < .0001) and polymicrobial infection (OR = 1.68; 95% CI, 1.22-2.32; P = .0014). Our data suggest that microbiological factors independently affect the outcome of nosocomial bloodstream infection.

Hospital epidemiology--1997

Selected issues for hospital epidemiology are presented. The issues of infection control, employee health and quality assessment cross the traditional department and reporting lines and require a multidisciplinary approach utilizing the epidemiological method. Programs seeking enhanced value--high quality outcomes per dollar cost--will likely support such an approach.

Infectious diseases consultation: impact on outcomes for hospitalized patients and results of a preliminary study

In the absence of detailed and specific data on outcomes, specialists are thought to contribute excessively to the cost of care. Infectious diseases physicians are at further risk of heavy utilization of medical resources because their patients often are those with serious complications or expensive underlying diseases (e.g., AIDS). By using readily available information in a sophisticated computer database, we examined the crude economic impact of an inpatient infectious diseases consultation and identified important matching variables for more refined analysis. The study was performed at LDS Hospital (Salt Lake City), a tertiary care facility with four full-time infectious diseases physicians. A total of 496 cases (patients who were seen by an infectious diseases consultant) were matched with 3,117 controls (patients who were not seen by an infectious diseases consultant). Matching was performed on the basis of age, sex, exact discharge diagnosis-related group, minimum length of hospital stay equal to the interval from admission to consultation for cases, and measures of the severity of illness (nursing acuity score and the number of secondary diagnoses). Cases had longer lengths of hospital stays, longer intensive care unit lengths of stays, and higher antibiotic costs than did matched controls, and if the consultation occurred in the last one-third of hospitalization, cases had shorter lengths of hospital stay and lower antibiotic costs than did controls.

Regionalization of tertiary care: impact of safe cardiovascular volumes in Iowa

The outcomes of cardiovascular care for individual patients are known to be better when treatment is provided in hospitals where the volume of such care is greater. We examined the impact of establishing safe volume thresholds on hospitals now performing such procedures in the state of Iowa by analysing the relationship between various volume thresholds and the number of hospitals that would continue to provide the service if the thresholds were enforced. Four procedures performed in hospitals in Iowa in 1990 were studied: coronary artery bypass graft surgery, vascular surgery, cardiac valve surgery, and cardiac catheterization. The analysis was conducted assuming current per capita procedure rates and repeated assuming reduced procedure rates. The study finds that the 12 hospitals now performing coronary artery bypass graft surgery in Iowa could be reduced to two, providers of vascular surgery could be reduced to as low as four from the current 33, the 10 providers of cardiac valve surgery might be reduced to one, and catheterization could be provided at two rather than 22 hospitals.

Identifying the hospitalized patient at risk for nosocomial bloodstream infection: a population-based study

Included in a 3-year population-based study were all patients (n = 64,281) admitted to a single tertiary care hospital (902 beds) using prospective hospital-wide surveillance for nosocomial infections. The objective of the study was to identify patients at risk for nosocomial bloodstream infection by using readily available hospital admission variables. After identifying potential risk factors for infection by univariate analyses, we derived multivariate models for predicting bloodstream infection by using logistical regression procedures. A total of 931 patients (1.45 per 100 admissions) developed a nosocomial bloodstream infection (2.2 episodes per 1000 patient-days) between 1 July 1987, and 30 June 1990. The crude mortality among infected patients was 34%, and the 319 deaths represented 22% of the total in-hospital mortality. Independent predictors of bloodstream infection were age, gender, primary diagnosis, and admission to a critical care unit. The sensitivity and specificity of the models for classifying patients as infected or noninfected were 81% and 81% for infants (1-11 months old) and 72% and 72% for adults, respectively. The negative predictive value of both models exceeded 99%. Applied to all patients on admission, the models we developed allowed us to survey only 28% of patients to identify more than 70% of those who will develop a nosocomial bloodstream infection.

Vancomycin-resistant enterococcal bacteremia: natural history and attributable mortality

Previous studies have shown that bacteremia due to vancomycin-resistant Enterococcus species (VRE) is associated with mortality of 17%-100%, but comorbid conditions may have confounded the estimates. We designed a historical cohort study to determine the mortality attributable to VRE bacteremia. Twenty-seven patients with VRE bacteremia were identified as cases. Within 7 days of the onset of bacteremia, severe sepsis developed in 12 patients (44%) and septic shock developed in 10 (37%). Case patients were closely matched to control patients without VRE bacteremia (1:1) by time of hospitalization, duration of exposure, underlying disease, age, gender, and surgical procedure. The mortality was 67% among cases and 30% among matched controls (P = 0.1). Thus, the mortality attributable to VRE bacteremia was 37% (95% confidence interval [CI], 10%-64%) and the risk ratio for death was 2.3 [CI, 1.2-4.1). We conclude that VRE bacteremia is associated with high rates of severe sepsis and septic shock. The attributable mortality approaches 40%, and patients who have VRE bacteremia are twice as likely to die than closely matched controls.

Comparative evaluation of an automated ribotyping system versus pulsed-field gel electrophoresis for epidemiological typing of clinical isolates of Escherichia coli and Pseudomonas aeruginosa from patients with recurrent gram-negative bacteremia

Ribotyping and macrorestriction analysis of chromosomal DNA using pulsed-field gel electrophoresis (PFGE) are among the more useful molecular epidemiologic typing methods. Because these techniques are labor intensive, automation of one or more steps may allow clinical laboratories to apply molecular typing methods. We compared the recently developed automated ribotyping system, the RiboPrinter Microbial Characterization System (DuPont), with PFGE as a means of typing clinical isolates of E. coli and P. aeruginosa. A total of 22 E. coli and 24 P. aeruginosa were typed by both PFGE and the RiboPrinter. When compared with PFGE typing of E. coli and P. aeruginosa, the RiboPrinter was less sensitive in identifying different strains, particularly among the isolates of P. aeruginosa. The RiboPrinter was completely automated and allowed up to 32 isolates to be typed within an 8-hour period. The pattern of results obtained in this study suggests that a heirarchical approach to molecular typing using the RiboPrinter Microbial Characterization System plus PFGE might be feasible. The RiboPrinter Microbial Characterization System promises to be a very useful addition to the expanding molecular typing armamentarium.

Vertical and horizontal transmission of unique Candida species to premature newborns

The number of nosocomial bloodstream infections due to Candida species in critically ill newborns is increasing. This pathogen may be vertically transmitted from the mother or nosocomially acquired in the nursery. The goal of this study was to identify the route of transmission of unique Candida species and strains from mothers to their preterm offspring. Specimens from mothers for fungal cultures were obtained before delivery, and specimens from infants for sequential fungal cultures were obtained at defined intervals. Candida species were identified by standard methods and were typed by electrophoretic karyotyping (EK) and restriction endonuclease analysis of genomic DNA (REAG) with pulsed-field gel electrophoresis. Antifungal susceptibility testing was performed on all isolates. Fungal cultures were positive for Candida species in 12 (63%) of 19 mothers' specimens and in seven (33%) of 21 infants' specimens. EK and REAG revealed that both the mother and the infant in three (14%) of 21 mother-infant pairs were colonized with the identical strain of Candida albicans. C. albicans was most commonly transmitted vertically. Candida parapsilosis colonized other infants and could not be accounted for by a maternal reservoir.

Predicting survival of patients with sepsis by use of regression and neural network models

OBJECTIVES

(1) To predict at the time of diagnosis of sepsis the subsequent occurrence of multiple organ failure and patient death; and (2) to compare the prediction accuracies of standard multiple logistic regression (MLR) and neural network (NN) models.

METHODS

The data were collected during a 5-year period for all patients (n=173) who met prospectively determined criteria for sepsis and had positive blood culture results while admitted in the surgical intensive care unit at the University Hospital of Geneva, Switzerland. These data formed the basis for a retrospective cohort study described elsewhere. The MLR model was adapted from existing data. An NN model of the feed-forward, back-propagation type was constructed for predicting the outcome of sepsis with bloodstream infection. Both models were constructed from randomly chosen subsets of patients and subsequently were evaluated on the remaining (independent) patients.

RESULTS

Survival after sepsis was predicted with an accuracy of 80% by the NN model, which used only information collected at the time of the diagnosis of sepsis. The development of multiple organ failure after the diagnosis of sepsis was predicted accurately (81.5%) with either the MLR or the NN model. Both the MLR and the NN methods depended on the interpretation of a likelihood quantity, requiring the choice of a threshold to make a survival prediction. The accuracy of the MLR models was very sensitive to the threshold value. The accuracy of the NN models was not sensitive to the choice of threshold, because they generated likelihood predictions that were distributed far from the middle range where the threshold was placed.

CONCLUSION

Compared with MLR models, the NN models were slightly more accurate and much less sensitive to the arbitrary threshold parameter.