Gram-negative Bacteremia In Patients Research Articles

Predictors and outcomes of multi-drug–resistant gram-negative bacteremia in patients with cancer: A retrospective cohort study at a tertiary cancer center in Oman

ObjectivesThis study aimed to delineate the characteristics and outcomes of gram-negative bacteremia (GNB) in oncology patients; analyze the risk factors for multi-drug–resistant (MDR) GNB; and assess its impact on the recurrence of bloodstream infection (BSI), hospital stay, and 30-day mortality. MethodsData, including demographics, clinical features, common cancers, and microbiologic findings, were collected retrospectively from electronic medical records of patients admitted with solid tumors and BSI episodes between January and December 2022. Fisher's exact tests were used to determine the effect of MDR-GNB on 30-day mortality and BSI recurrence. The Wilcoxon rank-sum test assessed the differences in the length of hospital stay. Logistic regression models identified the risk factors for MDR-GNB. ResultsAmong 1074 patients, 77 episodes of GNB bacteremia occurred in 59 individuals (47% male, median age 57.4 years). Of these, 37 (48%) were MDR-GNB. Carbapenem resistance was noted in 9.1% of GNB episodes. Previous antibiotic use was significantly associated with MDR-GNB (odds ratio 7.82; 95% confidence interval 2.52-24). MDR-GNB was linked to longer hospital stays (median 23 vs 10.5 days, P = 0.003) and higher recurrence rates than non-MDR-GNB (35.13% vs 5.0%, P <0.001). However, 30-day mortality did not significantly differ between the groups (35.14% vs 32.5%, P = 0.81). ConclusionPrevious antibiotic use predicted MDR-GNB in patients with solid tumor. MDR-GNB bacteremia increased the length of hospital stay and risk of recurrence compared with non-MDR-GNB bacteremia.

244. Predictors and Outcomes of Multi-Drug Resistant Gram-Negative Bacteremia in Patients with Solid Tumours

Abstract Background Multi-drug resistant Gram-negative bacteraemia (MDR-GNB) is a major concern in cancer patients. However, predictors and outcomes of MDR-GNB bacteraemia in patients with solid tumours are not well defined. The objective of this study was to identify the predictors and outcomes of MDR-GNB bacteraemia in patients with solid tumours. Methods This retrospective cohort study was conducted on patients with solid tumours admitted to a tertiary care hospital between January 2022 and December 2022. Demographic, clinical, and microbiological data were collected from electronic medical records. Univariate and multivariate analyses were performed to identify the predictors of MDR bacteraemia. The outcomes of interest included in-hospital 30-day mortality, length of hospital stay, and recurrence of bacteraemia. Results Among 1074 patients with solid tumours, a total of 77 episodes of Gram-negative bacteremias occurred in 59 patients (47% male) with a mean age of 57.4 years. Of which 37(48%) had MDR-GNB bacteraemia. The most common source of bacteraemia was biliary (34.2%) followed by central line-associated (21%) and abdomen(21%). Carbapenem resistance was documented in 7(9.1%) of all Gram-negative bacteremias of which 75% were harbouring NDM with or without OXA-48 among enterobacteralis. The predictors of MDR-GNB bacteraemia were prior use of antibiotics (odds ratio [OR] 7.82, 95% [CI] 2.52,24.2) and E.coli bacteremia (OR 6.56, 95% CI 1.26,34.22). The median length of hospital stay was significantly longer in the MDR group compared to the non-MDR group (23 vs. 10.5 days; p= 0.003) and recurrence of bacteremia was significantly more common in the MDR group (35.13% vs. 5.0%; p=&amp;lt; 0.001). The 30-day in-hospital mortality was not significantly different (35.14% vs. 32.5%; p=0.81). Conclusion This study identifies prior use of antibiotics and E.coli bacteremias as predictors of MDR-GNB bacteraemia in patients with solid tumours. Patients with MDR-GNB bacteraemia had longer hospital stays, and higher recurrence rates compared to those with non-MDR-GNB bacteremias. This highlights the importance of the judicious use of antibiotics and infection control measures to prevent MDR-GNB bacteraemia in this population. Disclosures All Authors: No reported disclosures

Evaluating antimicrobial duration for Gram-negative bacteremia in patients with neutropenia due to hematologic malignancy or hematopoietic stem cell transplantation.

In the management of Gram-negative bloodstream infection (GN-BSI), short antimicrobial courses have been increasingly demonstrated to be non-inferior to prolonged therapy, with lower risk of Clostridioides difficile infection (CDI) and emergence of multi-drug resistant (MDR) organisms. However, immunocompromised hosts were excluded from these studies. We investigated outcomes of short (≤10 days), intermediate (11-14 days), and prolonged (≥15 days) antimicrobial durations for GN-BSI in neutropenic patients. A retrospective cohort study was conducted on neutropenic patients with monomicrobial GN-BSI between 2018 and 2022. The primary outcome was a composite of all-cause mortality and microbiologic relapse within 90 days after therapy completion. The secondary outcome was a composite of 90-day CDI and development of MDR-GN bacteria. Cox regression analysis with propensity score (PS) adjustment was used to compare outcomes between the three groups. A total of 206 patients were classified into short (n = 67), intermediate (n = 81), or prolonged (n = 58) duration. Neutropenia was predominantly secondary to hematopoietic stem cell transplantation (48%) or hematologic malignancy (35%). The primary sources of infection included intra-abdominal (51%), vascular catheter (27%), and urinary (8%). Most patients received definitive therapy with cefepime or carbapenem. No significant difference in the primary composite endpoint was observed for intermediate versus short (PS-adjusted hazard ratio [aHR] 0.89; 95% confidence interval [95% CI] 0.39-2.03) or prolonged versus short therapy (PS-aHR 1.20; 95% CI 0.52-2.74). There was no significant difference in the secondary composite endpoint of CDI or MDR-GN emergence. Our data suggest that short antimicrobial courses had comparable 90-day outcomes as intermediate and prolonged regimens for GN-BSI among immunocompromised patients with neutropenia.

305. Prevalence of Ceftriaxone Susceptible, Piperacillin-tazobactam Non-susceptible Escherichia coli Bacteremia in Patients with Hematologic Malignancy

BackgroundPiperacillin-tazobactam (TZP) is common empiric and targeted therapy for gram-negative bacteremia in patients with hematologic malignancy. Resistance to TZP tends to occur concurrently with ceftriaxone (CRO) resistance; however, the prevalence of TZP nonsusceptibility in CRO susceptible Escherichia coli (E. coli) in patients with hematologic malignancy is unknown. Therefore, we sought to determine the prevalence of TZP nonsusceptible, CRO susceptible E. coli bacteremia at a cancer center.MethodsThis is a retrospective cohort study of adult (age > 18) patients with E. coli bacteremia admitted to the Leukemia or Stem Cell Transplant (SCT) services at The University of Texas MD Anderson Cancer Center (MDACC) between 8/2016 and 7/2019. Isolates were categorized according to current CLSI resistance breakpoints. A first isolate was defined as the first positive blood culture and subsequent episodes of bacteremia were defined as any E. coli isolate obtained at least 24 hours after the first negative blood culture.ResultsThe overall prevalence of TZP resistant CRO susceptible E. coli from 404 isolates was 7.7% and varied by service. There was a higher prevalence in the Leukemia service compared to SCT, 9.8% vs 2.5%, respectively (p < 0.01). 46% of isolates were CRO nonsusceptible, of which 91% were extended-spectrum beta-lactamase (ESBL) producers, identified by Vitek 2 or Accelerate Pheno. The TZP MIC50 was 4ug/ml, MIC90 was 128ug/ml, with an MIC range of 3ug/ml to ≥ 256ug/ml. The TZP MIC distribution varied based upon CRO phenotype. In CRO susceptible isolates the MIC50 and MIC90 were 4ug/ml and 64ug/ml, respectively, compared to 8ug/ml and 128ug/ml in CRO nonsusceptible isolates (p < 0.01). TZP resistance was more common in CRO nonsusceptible isolates (31.6% vs 12.0%, p < 0.01) and was more frequent with subsequent episodes of bacteremia compared to the first (39.5% vs 20.1%, p < 0.01).ConclusionIn patients with hematologic malignancy and E. coli bacteremia, TZP resistance is common with significant variations by CRO phenotype. TZP resistance becomes more common with subsequent episodes of bacteremia compared to the first. The clinical implications and genetic cause of this phenotype is currently unknown and warrants further investigation.Disclosures All Authors: No reported disclosures

116. Risk Factors and Clinical Outcomes of Carbapenem Non-Susceptible Gram-Negative Bacteremia in Patients with Acute Myelogenous Leukemia

BackgroundEarly administration of susceptible antibiotics is crucial in Gram-negative bacteremia (GNB), especially in immunocompromised patients. We aimed to explore risk factors and clinical outcomes of carbapenem non-susceptible (Carba-NS) GNB in patients with acute myelogenous leukemia (AML).MethodsCases of all GNB during induction or consolidation chemotherapy for AML in a 15-year period in a tertiary hospital were retrospectively reviewed. Independent risk factors for Carba-NS GNB were sought and its clinical outcomes were compared with those of carbapenem susceptible (Carba-S) GNB.ResultsAmong 485 GNB cases from 930 patients, 440 (91%) were Carba-S and 45 (9%) were Carba-NS GNB. Frequent Carba-NS isolates were Stenotrophomonas maltophilia (n = 23), Pseudomonas aeruginosa (n = 11), and Acinetobacter baumannii (n = 10). Independent risk factors for Carba-NS GNB were carbapenem use at the onset of GNB (aOR [95% CI], 78.6 [24.4–252.8]; P < 0.001), the isolation of imipenem-resistant A. baumannii in the prior 1 year (aOR [95% CI], 14.6 [2.7–79.9]; P = 0.002), time interval from chemotherapy to GNB ≥20 days (aOR [95% CI], 4.7 [1.7–13.1]; P = 0.003), and length of hospital stay ≥30 days (aOR [95% CI], 3.4 [1.3–9.1]; P = 0.013). Except breakthrough GNBs which occurred during carbapenem treatment, the frequency of Carba-NS GNB was 48% (19/40) in cases having ≥2 risk factors other than carbapenem use. 30-day overall mortality (Carba-NS, 36% vs. Carba-S, 6%; P < 0.001) and in-hospital mortality (Carba-NS, 47% vs. Carba-S, 9%; P < 0.001) were significantly higher in Carba-NS GNB.ConclusionCarba-NS GNB in AML patients was independently associated with the use of carbapenem, the past isolation of resistant organism, and late onset of GNB, and its clinical outcomes were poorer than those of Carba-S GNB. Carba-NS organisms should be considered for antibiotic selection in AML patients having these risk factors.Disclosures All authors: No reported disclosures.

Procalcitonin Is a Marker of Gram-Negative Bacteremia in Patients With Sepsis

BackgroundPrediction of the species of pathogen among patients with sepsis within hours would be helpful in accelerating proper treatment. As a potential method of shortening the time to identification, this study considered the usefulness of measuring procalcitonin (PCT) to predict blood culture (BC) results. MethodsThe authors retrospectively analyzed the data of patients with a diagnosis of sepsis in their hospital from December 2012 to December 2013. The authors analyzed all diagnostic episodes consisting of BC and PCT concentration. The diagnostic performance of PCT to predict gram-negative bacteremia was tested using a receiver operative characteristic curve. Logistic regression was constructed using the presence of gram-negative bacteria as the dependent variable. ResultsA total of 262 diagnostic episodes met the inclusion criteria. According to BC classifications, a significantly higher value of PCT was observed in bloodstream infections caused by gram-negative bacteria (26.7 ng/mL, 0.09–188.3) than that in bloodstream infections caused by gram-positive bacteria (0.84 ng/mL, 0.05–18.79) or Candida spp. (1.12 ng/mL, 0.07–49.68). A cutoff value of ≥3.39 ng/mL for PCT showed a sensitivity of 80%, a specificity of 71%, a positive predictive value of 35%, a negative predictive value of 91% and an area under the curve of 0.73 for gram-negative bacteremia identification by BC. Among the 122 diagnostic episodes with positive BC results, a cutoff value of ≥6.47 ng/mL for PCT yielded a sensitivity of 74%, a specificity of 81%, a positive predictive value of 82%, a negative predictive value of 75% and an area under the curve of 0.81 for gram-negative bacteremia identification. ConclusionsPCT may represent a useful tool for differentiating gram-positive from gram-negative bloodstream infection with a significantly higher PCT level indicating gram-negative bacteremia.

Bacteremic complications of intravascular catheter tip colonization with Gram-negative micro-organisms in patients without preceding bacteremia

Although Gram-negative micro-organisms are frequently associated with catheter-related bloodstream infections, the prognostic value and clinical implication of a positive catheter tip culture with Gram-negative micro-organisms without preceding bacteremia remains unclear. We determined the outcomes of patients with intravascular catheters colonized with these micro-organisms, without preceding positive blood cultures, and identified risk factors for the development of subsequent Gram-negative bacteremia. All patients with positive intravascular catheter tip cultures with Gram-negative micro-organisms at the University Medical Center, Utrecht, The Netherlands, between 2005 and 2009, were retrospectively studied. Patients with Gram-negative bacteremia within 48 h before catheter removal were excluded. The main outcome measure was bacteremia with Gram-negative micro-organisms. Other endpoints were length of the hospital stay, in-hospital mortality, secondary complications of Gram-negative bacteremia, and duration of intensive care admission. A total of 280 catheters from 248 patients were colonized with Gram-negative micro-organisms. Sixty-seven cases were excluded because of preceding positive blood cultures, leaving 213 catheter tips from 181 patients for analysis. In 40 (19%) cases, subsequent Gram-negative bacteremia developed. In multivariate analysis, arterial catheters were independently associated with subsequent Gram-negative bacteremia (odds ratio [OR] = 5.00, 95% confidence interval [CI]: 1.20–20.92), as was selective decontamination of the digestive tract (SDD) (OR = 2.47, 95% CI: 1.07–5.69). Gram-negative bacteremia in patients who received SDD was predominantly caused by cefotaxime (part of the SDD)-resistant organisms. Mortality was significantly higher in the group with subsequent Gram-negative bacteremia (35% versus 20%, OR = 2.12, 95% CI: 1.00–4.49). Patients with a catheter tip colonized with Gram-negative micro-organisms had a high chance of subsequent Gram-negative bacteremia from any cause. This may be clinically relevant, as starting antibiotic treatment pre-emptively in high-risk patients with Gram-negative micro-organisms cultured from arterial intravenous catheters may be beneficial.

Prediction of gram-negative bacteremia in patients with cancer and febrile neutropenia by means of interleukin-8 levels in serum: targeting empirical monotherapy versus combination therapy.

In a prospective observational study of 133 neutropenic episodes, interleukin (IL)-8 serum levels > 2000 pg/mL at the onset of fever had a sensitivity of 53% and a specificity of 97% as a predictor of gram-negative bacteremia (GNB; positive predictive value, 73%; negative predictive value, 94%). The rates of early death differed significantly between patients with high and those with low IL-8 levels (3/11 vs. 1/122; P< .01). Serum IL-8 levels at the onset of fever define a low-risk subgroup of patients who can safely be treated with monotherapy.

Concordance of endotoxemia with gram-negative bacteremia. A meta-analysis using receiver operating characteristic curves.

OBJECTIVE: To apply meta-analysis to compare the concordance between the results of 2 types of limulus amebocyte lysate (LAL) assay, gelation (GLAL) and chromogenic (CLAL), with the detection of gram-negative bacteremia in patients with suspected bacteremia. DESIGN: Meta-analysis using receiver operating characteristic-based analytical method. DATA SOURCES: MEDLINE literature search and manual reviews of article bibliographies together with direct approaches to authors of potentially eligible studies. STUDY SELECTION: The studies that were selected had all included at least 10 patients, of whom at least 2 patients were diagnosed with gram-negative bacteremia, and all had data available for extraction into a contingency table format. RESULTS: Fifty-six studies (28 GLAL and 28 CLAL studies) met the inclusion criteria. Studies were stratified by type of test (GLAL vs CLAL). Each analysis was repeated with smaller studies excluded. There was no difference between the 2 types of LAL assays. Among the CLAL studies, there was no difference between studies that did versus those that did not use the sepsis syndrome criteria as a basis for patient inclusion. Among 45 studies for which data on the proportion of non-Enterobacteriaceae were available, there was a trend toward higher concordance as this proportion increased. CONCLUSIONS: The concordance between the LAL test and the detection of gram-negative bacteremia in patients with suspected bacteremia is no higher with the CLAL assay than with the original GLAL version. However, the concordance is higher among studies with a higher proportion of non-Enterobacteriaceae among the gram-negative bacteremia isolates.

Gram-negative bacteremia in non-neutropenic patients: a 3-year review.

The causative organisms, clinical manifestations, factors influencing prognosis, and other epidemiological characteristics of 81 episodes of bacteremia due to gram-negative organisms, in non-neutropenic patients, were studied retrospectively during a 3-year period (1992-1994) at the Department of Internal Medicine of the University Hospital of Heraklion, Crete, Greece. The gram-negative bacteremia incidence was 2% and the overall mortality 12%. All 81 patients had fever; Escherichia coli was the most frequent organism isolated (from 47 patients--58%) and was associated with shock (9/47), disseminated intravascular coagulation (DIC) (8/47), anuria (5/47), adult respiratory distress syndrome (ARDS) (3/47), and pneumonia (1/47). Other less frequent gram-negative microorganisms were Klebsiella spp. (ten patients; 12%), Pseudomonas spp. (7; 7%), Salmonella spp. (5; 6%), Enterobacter spp. (5; 6%), Proteus spp. (3; 3.4%), Stenotrophomonas spp. (3; 3.4%), and Acinetobacter spp. (1; 1.2%). ARDS. shock, DIC, anuria, presence of central venous catheter, urinary catheter, unknown origin of infection and inappropriate treatment were significantly associated with a higher death rate. Early initiation of appropriate therapy was the most important intervention that favorably affected the outcome of gram-negative bacteremias in this patient population.

Elimination of an outbreak of gram-negative bacteremia in a hemodialysis unit

Background: The purpose of this study was to identify the cause of an unusual outbreak of gram-negative bacteremia in patients undergoing long-term hemodialysis. Methods: We performed direct observation and investigation of current dialysis techniques and facilities including microbiological sampling in a long-term hemodialysis unit in a tertiary care center. We also performed a retrospective review of medical charts and laboratory data of 10 patients undergoing long-term hemodialysis who experienced 11 episodes of gram-negative bacteremia between March and June 28, 1993. Results: All of these patients underwent dialysis by jugular venous access. Containers used to collect flush solution after priming of dialysis tubing remained unemptied for extended periods of time, and quantitative culture revealed more than 200 colony-forming units/ml gram-negative bacilli, including species isolated in blood cultures. Dialysis tubing and connector were left submerged in flush solution collection containers during priming, and the process of disinfecting tubing before patient connection had recently been discontinued. Control measures included emptying of flush containers after each use and daily decontamination. All dialysis tubing was to be disinfected before patient connection. Conclusion: Outbreak was due to contamination during dialysis setup. After institution of appropriate control measures, no new cases have occurred.

Concordance of endotoxemia with gram-negative bacteremia in patients with gram-negative sepsis: a meta-analysis.

The Limulus amebocyte lysate (LAL) assay is a sensitive method for detecting endotoxin. Using gram-negative (GN) bacteremia as the basis for comparison, concordance with endotoxemia in 45 studies could be expressed as an odds ratio. Calculation of summary odds ratios by the Mantel-Haenszel-Peto method indicated that the concordance of the results was no higher by the chromogenic LAL assay than by the gelation version, and the sensitivity was improved by only 11% (62 versus 51%). Endotoxemia was detected in 77 (68%) of 114 patients with bacteremia caused by an organism that was not a member of the family Enterobacteriaceae, whereas endotoxemia was detected in only 120 (45%) of 269 patients with bacteremia caused by a member of the family Enterobacteriaceae or an anaerobe (P < 0.001). This difference was also apparent for patients with GN bacteremia for whom a fatal outcome had been recorded. The prevalence of GN bacteremia in the tested population and the type of etiological agent are critical and previously unrecognized variables which affect the interpretation of the LAL test in patients with suspected sepsis.

Pseudomonas aeruginosa and Enterobacteriaceae bacteremia after biliary endoscopy: An outbreak investigation using DNA macrorestriction analysis

purpose: An outbreak of gram-negative bacteremia in patients undergoing endoscopic retrograde cholangiopancreatography (ERCP) was investigated to determine the sources of infection and to control transmission. patients, methods, and results: The incidence of post-ERCP bacteremia increased from 1.6% (60 of 3,696) procedures to 3.6% (53 of 1,454) procedures (relative risk 2.3, p <0.0001) after endoscopes were processed in a new automated disinfector. Bacteremia involved nine species of Pseudomonas and Enterobacteriaceae, which were also isolated from processed endoscopes. Seven epidemic strains with highly related genomic macrorestriction profiles each infected 2 or more patients, accounting for 29 (55%) episodes of post-ERCP bacteremia. Strains recovered from endoscopes and from the disinfector were associated with 22 (42%) and 5 (9%) bacteremic episodes respectively. Effective endoscope disinfection was achieved by cleansing and disinfection of a blind channel not processed in the disinfector, additional isopropanolair flush of all channels, and auto-disinfection of the disinfector. In the following period, the incidence of post-ERCP bacteremia returned to the pre-epidemic rate (1.7%, p = 0.0001). conclusion: Bacterial genome fingerprinting by macrorestriction analysis enabled delineation of a multi-strain outbreak of post-ERCP bacteremia. Cross-contamination, and to a lesser extent, common-source contamination, appeared related to inadequate disinfection of endoscopes processed in an automated disinfector.

Does a negative D-dimer exclude thrombosis?

The marked sensitivity of assays of cross-linked fibrin degradation products (XL-FDP) or D-dimer tests for the detection of physiologic fibrinolytic activity has been exploited to exclude the presence of an active thrombotic disorder. For example, as reviewed by Devine in this symposium, a lack of elevation of the concentration of XL-FDP has a negative predictive value of greater than 90% in patients with suspected deep venous thrombosis. Similarly, Deitcher and Eisenberg recently reported that a normal concentration of XL-FDP essentially excluded the presence of gram-negative bacteremia in patients with suspected sepsis.’ The negative predictive value of assays of XL-FDP in plasma from patients with these clinical conditions is dependent on the extent to which increases in intravascular fibrinolytic activity are reflected by increases in the concentration of XL-FDP. While assays of XL-FDP appear to be sensitive and specific for characterizing the intensity of physiologic fibrinolytic activity, it is important to recognize that differences in the immunogenicity of specific XL-FDP, as well as varying intensities of physiologic fibrinolytic response among individuals, may influence the extent to which a low or normal concentration of XL-FDP can be interpreted as excluding thrombosis or increased fibrinolysis.

Predictors of bacteremia and gram-negative bacteremia in patients with sepsis. The Veterans Affairs Systemic Sepsis Cooperative Study Group

We analyzed data from the Department of Veterans Affairs trial of steroid therapy for systemic sepsis to identify predictors of bacteremia and gram-negative bacteremia. Of the 2568 patients screened for entry in the trial, 465 met the following criteria: presence of four of seven clinical signs of sepsis; blood cultures at the time of screening; and complete data on nine clinical parameters. The multivariate logistic regression model was used to identify predictors of bacteremia and gram-negative bacteremia. Predicted probabilities of having these types of infections were calculated using the identified predictors. Patients were then classified into groups with and without bacteremia (and gram-negative bacteremia) based on the predicted probability. Misclassification error rates were calculated for each method of categorization by comparing the true with the predicted grouping of patients. Three factors were independently predictive of bacteremia and gram-negative bacteremia: elevated temperature, low systolic blood pressure, and low platelet count. Using these three factors, classification methods were identified that predicted blood infection better than chance, but misclassification was also high. For predicting bacteremia, the maximum predicted positive rate was 83%, with a specificity of nearly 100% and a sensitivity of only 5%. For predicting gram-negative bacteremia, the maximum predicted positive accuracy was 100%, with a specificity also of 100% and a sensitivity of almost 0%. Using simple clinical parameters, we could not predict either bacteremia or gram-negative bacteremia with sufficient accuracy to be clinically meaningful; however, our approach represents a step in the direction of forecasting the bacterial organism responsible for sepsis in advance of culture results.

Predictors of Bacteremia and Gram-negative Bacteremia in Patients With Sepsis

<h3>Background.——</h3> We analyzed data from the Department of Veterans Affairs trial of steroid therapy for systemic sepsis to identify predictors of bacteremia and gram-negative bacteremia. <h3>Methods. —</h3> Of the 2568 patients screened for entry in the trial, 465 met the following criteria: presence of four of seven clinical signs of sepsis; blood cultures at the time of screening; and complete data on nine clinical parameters. The multivariate logistic regression model was used to identify predictors of bacteremia and gram-negative bacteremia. Predicted probabilities of having these types of infections were calculated using the identified predictors. Patients were then classified into groups with and without bacteremia (and gram-negative bacteremia) based on the predicted probability. Misclassification error rates were calculated for each method of categorization by comparing the true with the predicted grouping of patients. <h3>Results.—</h3> Three factors were independently predictive of bacteremia and gram-negative bacteremia: elevated temperature, low systolic blood pressure, and low platelet count. Using these three factors, classification methods were identified that predicted blood infection better than chance, but misclassification was also high. For predicting bacteremia, the maximum predicted positive rate was 83%, with a specificity of nearly 100% and a sensitivity of only 5%. For predicting gram-negative bacteremia, the maximum predicted positive accuracy was 100%, with a specificity also of 100% and a sensitivity of almost 0%. <h3>Conclusions. —</h3> Using simple clinical parameters, we could not predict either bacteremia or gram-negative bacteremia with sufficient accuracy to be clinically meaningful; however, our approach represents a step in the direction of forecasting the bacterial organism responsible for sepsis in advance of culture results. (<i>Arch Intern Med</i>. 1992;152:529-535)

Bacterial translocation and gram-negative bacteremia in patients with hematological malignancies.

Between 1976 and 1982, Enterobacteriaceae and Pseudomonas aeruginosa were prospectively counted in fecal specimens from leukemic patients with gram-negative bacteremia. The strains isolated from the blood and feces of 55 patients were compared. Translocation of the dominant fecal strain of Enterobacteriaceae or P. aeruginosa was observed in 45 cases (82%) and was strongly associated with granulocytopenia of less than 10(2) cells/microliter (P less than .0001). Thirteen (81%) of 16 patients with bacteremia caused by P. aeruginosa were intestinal carriers of the same strain, whereas only 2 (5%) of 39 patients with bacteremia caused by Enterobacteriaceae were carriers of P. aeruginosa. Bacterial translocation of Enterobacteriaceae was not associated with an abnormally high fecal population of the translocating strain. Prospective quantitative and qualitative analyses of fecal flora were useful in forecasting the most probable translocating gram-negative organism in neutropenic leukemic patients with clinical signs of bacteremia.