Incidence and clinical significance of positive blood cultures in febrile episodes of patients with hematological malignancies

Impact of multidrug resistance on outcomes in hematologic cancer patients with bacterial bloodstream infections

Despite the improved outcomes in patients with hematological malignancies, infections caused by multidrug-resistant organisms (MDROs) pose a new threat to these patients. We retrospectively reviewed the patients with hematological cancer and bacterial bloodstream infections (BSIs) at a tertiary hospital between 2003 and 2022 to assess the impact of MDROs on outcomes. Among 328 BSIs, 81 (24.7%) were caused by MDROs. MDRO rates increased from 10.3% (2003-2007) to 39.7% (2018-2022) (P < 0.001). The 30-day mortality rate was 25.0%, which was significantly higher in MDRO-infected patients than in non-MDRO-infected patients (48.1 vs. 17.4%; P < 0.001). The observed trend was more pronounced in patients with newly diagnosed diseases and relapsed/refractory disease but less prominent in patients in complete remission. Among MDROs, carbapenem-resistant Gram-negative bacteria exhibited the highest mortality, followed by vancomycin-resistant enterococci, methicillin-resistant Staphylococcus aureus, and extended-spectrum β-lactamase-producing Enterobacteriaceae. Multivariate analysis identified independent risk factors for 30-day mortality as age ≥ 65 years, newly diagnosed disease, relapsed/refractory disease, MDROs, polymicrobial infection, CRP ≥ 20 mg/L, and inappropriate initial antibiotic therapy. In conclusion, MDROs contribute to adverse outcomes in patients with hematological cancer and bacterial BSIs, with effects varying based on the underlying disease status and causative pathogens. Appropriate initial antibiotic therapy may improve patient outcomes.

How I manage haematology patients with septic shock

Patients with a variety of haematological conditions are at risk of infection and its most serious complication: septic shock. Mortality for septic shock remains high and especially so in patients with haematological malignancy and following bone marrow transplantation. However, advances in the treatment of severe sepsis have improved mortality rates even though evidence for the management of severe sepsis in haematology patients is limited. Wherever possible this review will concentrate on evidence directly applicable to haematology patients but inevitably will have to extrapolate evidence from other patient groups. The Surviving Sepsis Guidelines 2008 provide information on best practice in the management of patients with severe sepsis and septic shock and are broadly applicable though not specific to haematology patients. This review summarizes a practical approach to the management of severe sepsis in haematology patients and highlights areas of research which may bring new treatments in the future. The review is limited to the management and initial resuscitation of septic shock in adult haematology patients and will not address the detailed intensive care management of these patients or the management of severe sepsis in children.

Infections in childhood acute lymphoblastic leukemia: an analysis of 222 febrile neutropenic episodes

A retrospective analysis was performed on febrile neutropenic episodes in patients with acute lymphoblastic leukemia (ALL) from 1992 to 2002. There were 222 febrile neutropenic episodes in 266 ALL patients with documented ANC < 500/mm(3). Of the 222 episodes, 98 (44%) had documented focus of infection; the rest were fever without focus. There were 274 different sites of infection in the 98 episodes of documented focus of infection; pulmonary infections were the commonest site of infection (27.3%) followed by HEENT (22.9%). Of 69 bacterial isolates, gram-negative bacteria (n = 46, 67%) were twice as common as gram-positive bacteria (n = 23, 33%). Most common site of isolation for gram-negative bacteria was blood (50%) followed by urine (32.6%). Blood (78.3%) was predominant site of isolation of gram-positive bacteria followed by HEENT (8.7%). Escherichia coli (45.7%) was the commonest gram-negative isolate, while Staphylococcus aureus (39%) was the commonest gram-positive bacterial isolate. There were a total of 22 fungal isolates, the majority from urine (n = 12) and HEENT (n = 9). Of the 22 fungal isolates, 19 were detected in induction phase of chemotherapy. A total of 95/222 (42.8%) febrile neutropenic episodes improved with first-line antibiotic therapy, while modification was required in 127 episodes (57.2%). Antifungal therapy was used in 86 episodes (38.7%). There were a total of 13 deaths, 6 each during induction and intensification/consolidation phases, while 1 died during maintenance phase. Of the 13 deaths, 10 had pneumonia, 8 had bacteremia, and 7 had fungal infection. The current study stresses the importance of frequent reviewing of type, frequency, severity, and outcome of infection complications over the years to detect changing epidemiological patterns. The majority of fungal infections were detected during induction chemotherapy, which highlights the need to consider this type of infection in the evaluation of patients.

Expression of MxA protein in blood lymphocytes of children receiving anticancer chemotherapy

The aim of the study was to evaluate whether IFN-alpha/beta-inducible MxA protein expression in children receiving anticancer treatment can be used as an indicator for virus infections during the febrile episodes. Twenty-six children with mainly hematological malignancies entered the study. Children with laboratory-confirmed virus infections had clearly elevated MxA protein levels compared to their counterparts with bacterial or unknown etiology. MxA protein expression increased moderately following the administration of cytostatic agents, even though these children had no clinical signs of infection.

Risk of bacteraemia and mortality in patients with haematological malignancies

This study examined the association between type of haematological malignancy, risk of bacteraemia and risk of mortality, with emphasis on the impact of bacteraemia type on mortality. A population-based cohort design was used, and all patients aged > or = 15 years with an incident haematological malignancy who were living in North Jutland County, Denmark, during 1992-2002 were included in the study. Among 1666 patients with an incident haematological malignancy, 358 (21%) suffered an episode of bacteraemia during a median follow-up period of 1.1 years (quartile 0.2-3.4) from the date of cancer diagnosis (overall incidence rate of 96/1000 person-years). In comparison to Hodgkin's disease, adjusted incidence rate ratios (IRRs) were 23.3 (95% CI, 10.0-54.5) for acute myeloid leukaemia, 3.8 (95% CI, 1.5-9.3) for multiple myeloma, and 2.2 (95% CI, 0.9-5.1) for non-Hodgkin's lymphoma or chronic lymphatic leukaemia. Overall cumulative 30-day mortality was 32% (95% CI, 27-37), and 90-day mortality was 50% (95% CI, 44-55). In comparison with acute myeloid leukaemia, adjusted mortality rate ratios (MRRs) were close to 1.0 for other haematological malignancies. In comparison to bacteraemia caused by Gram-positive bacteria, adjusted MRRs were 1.0 (95% CI, 0.6-1.5) for Gram-negative bacteraemia, and 1.9 (95% CI, 1.1-3.3) for polymicrobial bacteraemia or fungaemia. Thus, the risk of bacteraemia varied greatly according to the type of malignancy, while mortality rates were similar for these diseases, although dependent on the type of bacteraemia. Polymicrobial bacteraemia or fungaemia was associated with higher mortality.

Assessment of systemic inflammation markers to differentiate a stable from a deteriorating clinical course in patients with febrile neutropenia

In this study, we evaluated the predictive values of procalcitonin (PCT), C-reactive protein (CRP), interleukin-6 (IL-6) and serum amyloid A (SAA) for determining the clinical course in febrile neutropenic patients. Daily plasma analyses during the fever course were performed in 101 episodes with fever and chemotherapy-induced neutropenia (neutrophil count <0.5 x 10(9)/L). Procalcitonin (PCT) and IL-6 values were significantly higher in febrile episodes in patients who developed complications. Procalcitonin with a cut-off value of < or =0.4 ng/mL or IL-6 < or =50 pg/mL 3 d after fever onset indicated daily high negative predictive values (NPVs) (91-100%) for episodes with complications. No marker could predict deterioration; however, daily low plasma concentrations of PCT or IL-6 during the first 8 d of fever were found to be a good predictor of no subsequent complications in neutropenic patients and therefore to be a helpful tool for limiting anti-microbial therapy.

Use of inflammatory markers for early detection of bacteraemia in patients with febrile neutropenia

The aim of the study was to evaluate the ability of procalcitonin, C-reactive protein, serum amyloid A, interleukin-6 and interleukin-8 to predict bacteraemia during the 2 first d of fever in neutropenic patients. A total of 94 febrile neutropenic episodes in 60 patients were studied. Plasma samples were analysed at 10-h intervals from the onset of fever. Clinical events were categorized into 4 groups: 1) bacteraemia caused by other agents than coagulase-negative staphylococci (non-CNS bacteraemia) (n = 21), 2) coagulase-negative staphylococci bacteraemia (n = 15), 3) microbiologically or clinically documented infection without bacteraemia (n = 26) and 4) fever of unknown origin (n = 32). In non-CNS bacteraemia all markers, except for serum amyloid A, showed significantly higher levels compared to patients with fever of unknown origin (p < 0.05). For non-CNS bacteraemia the highest negative predictive value was found for procalcitonin (94%), followed by interleukin-6 (89%), C-reactive protein (88%) and interleukin-8 (87%). Procalcitonin, with a cut-off level of 1.4 ng/ml during 10-20 h after fever onset, showed the highest positive predictive value (67%) for a non-CNS bacteraemia. In conclusion, the value of the analysed markers to predict a non-CNS bacteraemia in neutropenic patients was limited due to low sensitivity and positive predictive value. However, procalcitonin, interleukin-6, C-reactive protein, and interleukin-8 could give useful information for the clinician in excluding a non-CNS bacteraemia.

[Etiology of bacterial infections in febrile neutropenic patients: the role of the laboratory in the diagnosis]

EPIDEMIOLOGICAL EVOLUTION: Until the mid-eighties, infectious complications (pneumonia, septicemia) observed in neutropenic patients were, in 70% of cases, of bacterial origin with Gram negative bacillae (Escherichia coli, Klebsiella sp, Pseudomonas aeruginosa) isolated 8 times out of 10. Among the Gram positive bacteria, Staphylococcus aureus predominated. The etiological profile of bacterial infections has since evolved with a predominance (60 to 70%) of Gram positive bacteria (coagulase-negative staphylococci, viridans streptococci) and a change in the epidemiology of the Gram positive bacteria notably with a lesser frequency of P. aeruginosa infections. THE GRAM POSITIVE BACTERIA: Coagulase-negative staphylococci are among the first germs responsible for nosocomial bacteremia (central venous catheters) and they are usually multiresistant. Viridans streptococci are a frequent cause of bacteremia; they are generally sensitive to antibiotics active on Gram positive bacteria, but the incidence of resistant strains is increasing. Enterococci are in majority responsible for colonisation in neutropenic patients and less frequently for infections; they raise the problem of resistance to antibiotics, notably to glycopeptides. Other Gram positive bacteria can be responsible for infections in neutropenic patients; it is crucial that they be identified because they require treatment with an appropriate antibiotic. GRAM NEGATIVE BACTERIA: Among the enterobacteria, Escherichia coli is predominantly isolated and raises the problem of the increasing incidence of resistance to fluoroquinolone. Pseudomonas aeruginosa, less frequently responsible today, remains associated with a far greater rate of mortality than that observed with the other microorganisms. Other Gram negative bacteria can be identified; they require an adapted antibiotherapy because they are often naturally multiresistant to antibiotics. THE ROLE OF THE LABORATORY: For the diagnosis of infections in neutropenic patients, the microbiology laboratory has a determinating role. The laboratory ensures the analysis of various biological examinations: blood cultures, methods permitting the diagnosis of an infection on a permanent catheter, copro-cultures (research for common enteropathogens, quantification in the case of digestive decontamination, screening for multiresistant bacteria), cytobacteriological examination of urine, samples of respiratory origin, cytobacteriological examination of cerebro-spinal fluid...).

[Incidence of bacteremia among histologic groups of hematologic malignancies]

INTRODUCTION

Subgroups having dissimilar prognoses are being identified among cancer patients with infection. Previous studies have suggested that these differences may be related to the histologic diagnosis, but this issue has not as yet been demonstrated.

METHODS

We reviewed the medical records of all patients admitted with acute leukemia (AL) or lymphoma (ML) from 1988 to 1998. Incidence of bacteremia was calculated for the following subgroups: acute lymphocytic leukemia (ALL), acute myelocytic leukemia (AML), AML following refractory anaemia with excess blasts (AML-RAEB), high-grade ML (HGML), intermediate-grade ML (IGML), low-grade ML (LGML) and indeterminate ML (IML). Kaplan-Meier curves of time to the first positive blood culture were constructed and compared by means of log-rank test.

RESULTS

In the period covered there were 244 new diagnoses of AL or ML: 62 AML, 32 ALL, 20 AML-RAEB, 78 HGML, 7 IGML, 37 LGML and 6 IML. At the end of the study period, 44 patients were alive, 147 were known to have died at a certain date and 53 had been formally lost to follow-up (most of them, transferred for hospice care). Among 684 blood cultures, there were 51 contaminations and 155 significant isolates. Among the latter, gram-positive bacteria were isolated in 74 and gram-negative bacteria in 47; in 27 cases more than 1 bacterial species were recovered. Fungi were isolated in 7 cases. The incidence of bacteremia expressed as cases per 1000 patient-days was 5.80 for AML, 5.03 for AML-RAEB, 1.56 for ALL, 0.21 for HGML and 0.40 for the remaining ML. Time to the first positive blood culture was significantly shorter for AML than for any other group, and was shorted for ALL and AML-RAEB than for ML.

CONCLUSION

Differences in the incidence of bacteremia were observed among histologically-defined groups of unselected patients with hematologic malignancies.

Increased incidence of bacteraemia due to viridans streptococci in an unselected population of patients with acute myeloid leukaemia

The aetiology, clinical characteristics and outcome of bacteraemia in patients with acute myeloid leukaemia were studied. All positive blood cultures collected at a haematological ward during 2 7-y periods were evaluated. Altogether, 274 episodes of bacteraemia in 152 patients were recorded, 80 episodes during 1980-86 and 194 during 1990-96. During the 2 periods, trimethoprim-sulfamethoxazol in combination with amikacin was the first-line empirical therapy in patients with neutropaenia and fever. In 1990, antimicrobial prophylaxis with ciprofloxacin and fluconazole was introduced. The incidence of bacteraemia due to viridans streptococci or coagulase-negative staphylococci increased from the first period to the second, whereas the incidence of Enterobacteriaceae decreased. In granulocytopaenic patients during 1990-96, viridans streptococci accounted for 21% of the isolates and in patients treated prophylactically with fluoroquinolone, viridans streptococci accounted for 31%. All viridans streptococci were sensitive to penicillin. At the time of the positive blood cultures, the patients of the second period were granulocytopaenic in 83% of the episodes. The mortality related to septicaemia during the later period was 13% and only 1 of 33 (3%) of the patients with viridans streptococci died. Eight patients (9%) died in relation to septicaemia following curative antileukaemic therapy.

Infectious complications in breast cancer patients undergoing peripheral blood stem cell transplantation: a single center retrospective analysis towards outpatient strategy

Infectious complications were retrospectively analyzed in 129 transplants, performed in 90 patients, to identify characteristics that qualify breast cancer patients for outpatient-based PBSCT. Thirty-one cases (24%) did not develop fever. Of the remaining 98 cases, 84.7% developed fever during severe neutropenia. On univariate analysis, disease stages II-III, first PBSCT, mucositis grades II-IV and the use of two alkylators were associated with a higher risk of fever development. The latter two factors also affected fever occurrence on multivariate analysis. A longer median time to fever onset was observed in patients conditioned with single as compared to double alkylating agent-containing regimens (respectively 8th vs 6th day, P < 0.00001). As compared with metastatic breast cancer (MBC), high risk breast cancer showed a 2.3-fold increased risk of developing early fever during neutropenia (CI 2.3-3.8), remaining the only variable still significant on multivariate analysis (P = 0.0039). Combination antibiotic therapy was equivalent to single agent therapy. Patients suffering from microbiologically documented fever were at higher risk of undergoing second-line antibiotic therapy. In conclusion, MBC patients treated with a conditioning regimen containing only one alkylating agent and adequate prophylaxis for mucositis may qualify for outpatient-based PBSCT on the basis of a lower risk of infection.

Utility of obtaining blood cultures in febrile neutropenic patients undergoing bone marrow transplantation

Infection remains an important cause of morbidity and mortality after bone marrow or stem cell transplantation. To evaluate the role of obtaining blood cultures for intermittent or persistent fever in neutropenic patients on antibiotic therapy, we performed a retrospective chart review of 196 consecutive patients admitted to the Bone Marrow Transplant Unit at the University of North Carolina Hospitals from 1995 to 1998. From the cohort of 196 patients, 154 patients developed neutropenic fever. The initial blood culture was positive in 16 of 145 patients during the first fever episode giving a prevalence of 11%. From the total of 109 patients that had blood cultures drawn after day 1 of fever, five patients had blood cultures positive for a pathogen, a prevalence of 4.6%. In only one patient, did blood cultures drawn after day 1 identify an organism not present on day 1 (prevalence 0.9%). After reviewing the results in the first 105 patients, we changed our timing of collection of blood cultures. Forty-nine patients were treated in this manner and we found that the mean number of blood cultures decreased from 9.2 to 4.7 per patient without a change in the frequency of infectious complications or length of hospitalization.

Effects of anti-neoplastic agents on the recovery of bacteria and yeasts in an automated blood culture system

The effects of 7 anti-neoplastic agents on the recovery of 5 aerobic gram-positive cocci, gram-negative rods and yeasts were studied in 2 different automated blood culture systems using an experimental model. In the absence of anti-neoplastic agents, the growth of gram-positive cocci was detected significantly earlier in standard than in FAN aerobic bottles. In the presence of 100 microM doxorubicin, however, the growth of gram-positive cocci was totally inhibited in standard culture conditions, while in FAN bottles the agent has no inhibitory effect. Etoposide at a concentration of 100 micromol/l also significantly delayed the growth of cocci in standard conditions. Neither the culture bottles nor the anti-neoplastic agents tested had any effect on the growth of gram-negative rods and yeasts. The results suggest that the anti-neoplastic agents present in blood might disturb the growth of gram-positive cocci in blood culture. This should be considered when validating blood culture systems and evaluating blood cultures of chemotherapy-receiving febrile patients.

Continuous-monitoring blood culture screening system improves the detection of bacteremia in neutropenic patients

One reason for the underdiagnosis of bacteremia as an etiology of fever in neutropenic patients might lie in inefficient blood culture techniques. This possibility was investigated in a retrospective study by comparing the efficacies of a manual and an automated blood culture system used to detect bacteremia in such patients. All neutropenic fever episodes accompanying all 93 intensive chemotherapy cycles of 26 consecutive patients with acute myelogenous leukemia (AML) enrolled in Finnish Leukemia Group AML 86 trial and treated in Oulu University Hospital over 3 1/2 years were analyzed. The chemotherapy protocol and the supportive care of the patients remained the same during the whole period. In 1990-91 the blood cultures were made manually and in 1992-93 with an automated continuous-monitoring culture screening system. Evaluable febrile episodes numbered 53 during 1990-91 and 73 during 1992-93. There was a statistically significant increase (p < 0.05) of culture-positive episodes, from 21% to 40%, when the continuous-monitoring system was adopted. The new method proved to be better than the manual one in detecting bacteremia of neutropenic patients.

An autopsy study of systemic fungal infections in patients with hematologic malignancies

The aim of this study was to determine the incidence of fungal infections detected on autopsy in a group of 40 patients with hematologic malignancies treated with intensive chemotherapy or bone marrow transplantation, and to evaluate the risk factors for fungal infections. A control group included 38 patients with nonhematologic diseases and without granulocytopenia but with at least one of the known risk factors for fungal infections. Standard histopathological and microbiological methods were used. A higher incidence of invasive fungal infections was found in patients with hematologic malignancies as compared to the control group (p < 0.01). The predominant causes of fungal infections were Candida albicans and Aspergillus spp. The incidence of fungal infections caused by Aspergillus was higher (p < 0.05) in patients with hematologic malignancies than in the control group. The independent risk factors for fungal infections were fungal colonization, number of antibiotics and duration of antibiotic therapy, duration of fever and skin rash. A higher proportion of fungal infections was diagnosed on autopsy than during the patients' life (p < 0.01).

Infections in neutropenic patients I: aetiology

Improvement in supportive care including the introduction of new antibiotics, antiviral and antifungal agents and haematopoietic growth factors have all contributed to a decreased chemotherapy-related mortality and morbidity in cancer patients. However, infections/septic shock during neutropenia still constitutes a major threat to these patients. Most patients develop fever during neutropenia and in 20-40% a manifest bacteremia is documented. In patients with prolonged neutropenia, the risk for fungal infections is increased. The spectrum of bacterial, fungal and viral infections in the neutropenic patient is reviewed.