Bloodstream infections in cancer patients with febrile neutropenia

Abstract

Bloodstream infections (bacteraemia) account for approximately 25–30% of febrile episodes in patients with febrile neutropenia (FN). In developed countries, Gram-positive pathogens predominate. Mortality is higher in Gram-negative bacteraemia. A recent study involving 2142 patients with FN was reviewed, including 168 patients with Gram-negative bacteraemia (mortality 18%), 283 patients with Gram-positive bacteraemia (mortality 5%) and 48 patients with polymicrobial bacteraemia (mortality 13%). Among patients who received prophylactic antibiotics, Gram-positive bacteraemia was far more common than Gram-negative bacteraemia (75% vs. 25%), compared with approximately 50% of each in patients without prophylactic antibiotics. Patients with a Multinational Association for Supportive Care in Cancer (MASCC) score <15 had a 36% mortality compared with 3% if the MASCC score was >21. The MASCC score may help risk stratification of patients with FN and bacteraemia, although these data require confirmation. In two series of patients from developing countries (Lebanon and Malaysia), Gram-negative bacteraemia was more common and mortality was higher. In developing countries, Gram-negative bacteraemia may be more frequent due to less use of prophylactic antibiotics and central lines. Laboratory markers may have predictive and prognostic value for bacteraemia in patients at the onset of FN, including mannose-binding lectin, interleukin (IL)-6, IL-8 and procalcitonin, but further studies are required before they can be recommended. New therapies are required to lower the mortality in patients with FN with a high risk for bacteraemia.

Introduction

Infection is still a frequent and serious problem in cancer patients treated with chemotherapy who then become neutropenic. Despite major advances in the therapy of febrile neutropenia (FN), including more effective and less toxic empirical broad-spectrum antibiotics that can be used as outpatient therapy in selected low-risk patients, availability of better empirical antifungal regimens in patients who do not respond to standard antibiotics and increased use of standard and long-acting granulocyte colony-stimulating factors, FN still remains a very difficult therapeutic problem with a significant mortality rate. The most severe bacterial infections in these patients are bloodstream infections (bacteraemia) with or without a site of primary infection.

Factors predictive of bacteraemia in patients with FN have been studied in the past [1] but the model developed was not very helpful. Bacteraemia during FN has mostly been studied in patients with haematological malignancies, especially those receiving haematopoietic stem cell transplants [2], [3]. In those series, the risk of developing bacteraemia varied with age, type of malignancy and type of transplant. There are fewer data for solid tumours, but neutropenia is an important predisposing factor [4]. The present-day situation for bacteraemia in the FN population is the subject of this manuscript.

Bacteraemia rates in large trials testing empirical treatments for FN remain high, as evidenced by a recent co-operative group publication by Viscoli et al. [5] in which 29% of FN episodes were associated with bacteraemia. In that study, the pathogens causing bacteraemia were approximately equally divided among Gram-positive and Gram-negative organisms. Mortality in patients with bacteraemia is high, especially when a clinical site of infection is identified [6], which is designated as ‘complex bacteraemia’. Although bacteraemia rates may be a little lower, they are still frequent. The Infectious Diseases Group of the European Organization for Research and Treatment of Cancer (EORTC-IDG) trials had a 32% rate of bacteraemia in 1973, which dropped to 22% in 1994 [7]. However, their experience was that the proportion of Gram-positive and Gram-negative pathogens causing bacteraemia changed over this time period. The proportion of Gram-negative and Gram-positive pathogens went from 71% and 29%, respectively, in 1973–1978 to 33% and 67%, respectively, in 1992–1994 [7]. This observation has been confirmed by others [8].

The reason for the increase in Gram-positive pathogens is probably multifactorial. The main theoretical cause is the increased use of prophylactic antibiotics such as the fluoroquinolones [9], which tend to lower the number of Gram-negative pathogens, thus the proportion of Gram-positive pathogens may then increase. Other possible causes include the increased use of central venous catheters in these patients, the chosen chemotherapy which may cause significant mucositis, and the increased use of antacids, which may all select for Gram-positive microorganisms. This change has mainly been seen in developed countries that can better afford prophylactic antibiotics and the use of central catheters, which will be discussed further.

The main Gram-negative organisms isolated that cause bacteraemia remain Escherichia coli, Klebsiella sp. and Pseudomonas aeruginosa, as has been the pattern for many years. Other Gram-negative organisms are also seen but far less often. Coagulase-negative staphylococci (CoNS) remain the main Gram-positive pathogens causing bacteraemia, but again a variety of other Gram-positive organisms are also isolated less frequently.

A few selected recent publications describing series of patients with FN will be reviewed to see whether the epidemiology and outcomes of treatment are changing and at the same time to see whether there are any good methods to predict which patients with FN will develop bacteraemia. We may then be able to develop new approaches to improve the results of therapy in this seriously ill patient population.