Management of Infections in Patients With Acute Leukemia

Introduction

Almost a decade has passed since our review on the management of infectious problems encountered in the care of patients with acute leukemia was published in Oncology.[1] In the interim, several studies have been published on that subject, which, while contributing important new insights, have also served to emphasize the fundamental principles of management first formulated and proven in the setting of acute leukemia and intensive antileukemia treatment almost 30 years ago (Table 1). These principles have informed the management of infections in not only the neutropenic state but also the state of profound immune suppression, as exemplified by graft-vs-host disease or the acquired immune deficiency syndrome.

This article will describe recent developments relating to the management of infectious illnesses in patients rendered neutropenic by leukemia or its treatment. The discussion will focus on the increasing repertoire of available antimicrobial drugs and adjunctive agents, viewed against the background of newly emerging resistant organisms and special problems, such as the increased use of indwelling venous catheters.

Modern treatment approaches to neutropenic fever are all predicated on the principle of prompt empiric initiation of broad-spectrum antibiotics directed against potential pathogens, even in the absence of a localizing inflammatory reaction or laboratory-based documentation of infection. Indeed, this practice has been so effective in reducing mortality that comparative studies that utilize mortality as an end point currently require very large numbers of patients—usually several hundred per arm.[2]

Smaller studies that do not have sufficient power to detect mortality differences have therefore focused on such end points as the duration of fever, hospital stay, and treatment-related costs. In the United States, this trend is driven, in part, by economic pressures imposed by the recent cost-containment trends in delivery in general.

Although several meta-analyses have been performed to address the issue of empiric antibiotic therapy, many are subject to the problems inherent in the source studies, as well as the usual criticisms relating to study selection and bias.[3,4] A particularly difficult problem is the publication bias against studies with negative results. All of these factors complicate the ability to in-terpret the accumulated literature on the subject of infections in the neutropenic host.

Anticipating some of these problems, the Immunocompromised Host Society convened a consensus panel in 1988 to formulate guidelines for the conduct and reporting of clinical trials in patients with neutropenic fever. The report of that group, issued in 1990,[5] serves as a useful benchmark for the evaluation of clinical trials in this area.

Antibacterial Approaches

Prophylaxis Against Gram-Negative Infections

The gastrointestinal (GI) tract is a major source of bacteria in patients who develop clinical or subclinical mucositis as a result of cytotoxic therapy. Common organisms include Escherichia coli, Klebsiella species, and Pseudomonas aeruginosa.

Oral GI prophylaxis has been advocated to combat the dissemination of these organisms early in the course of profound cytotoxic drug-induced neutropenia and to inhibit late-onset GI colonization with drug-resistant pathogens. Such prophylaxis may be more important in patients expected to have prolonged neutropenia or repetitive cycles of myelosuppressive therapy, particularly if this is associated with severe chemotherapy-induced mucositis.

Early studies using trimethoprim(Drug information on trimethoprim)-sulfamethoxazole,[6,7] as well as more recent studies using systemic quinolones,[8] showed significant reductions in the incidence of gram-negative infections, especially in patients receiving more highly cytotoxic therapy and those with more severe aplasia. However, the impact on other important outcomes was limited in these studies, and prophylactic treatment with trimetho-prim-sulfamethoxazole was found in some studies[6] to be associated with adverse events, including the emergence of resistant bacteria and fungal infections.

Although the emergence of resistance with systemic fluoroquinolone usage remains a concern, a meta-analysis of several trials in which these agents had been used prophylactically concluded that the risk was low.[9] Trials comparing fluoroquinolones to trimeth-oprim-sulfamethoxazole suggest greater efficacy for the fluoroquinolones in preventing early-onset bacterial infections but little effect on other end points.[9-11] However, oral norfloxacin(Drug information on norfloxacin) (Noroxin), which is not absorbed but reaches high concentrations in the gut lumen, appears to spare anaerobic organisms and thus maintain colonization resistance in the GI tract against fungal overgrowth or the acquisition of new aerobic pathogens throughout the period of prolonged profound chemotherapy-induced neutropenia.[11-14]

Empiric Treatment of First Fever

Fever in the neutropenic host must be interpreted as a sign of infection, even in the absence of other localizing or systemic symptoms, and must prompt the empiric institution of broad-spectrum antibiotics. Numerous trials continue to validate this concept using agents directed against aerobic and facultative gram-negative bacteria, especially P aeruginosa.

To this end, the non–cross-resistant combination of an aminoglycoside and an antipseudomonal penicillin provides complementary mechanisms of action and potential antibacterial synergy. However, the renal, auditory, and vestibular toxicity that may be associated with prolonged aminoglycoside use is substantial. Therapeutic drug level monitoring, which is necessary to limit aminoglycoside toxicity, adds to the cost of therapy. Many recent studies, therefore, have focused on the identification of non–aminoglycoside-containing regimens, with a distinct trend toward the development of single-agent therapies.

Several beta-lactam and carbapenem antibiotics (eg, aztreonam [Azactam], cefepime [Maxipime], ceftazidime(Drug information on ceftazidime), imipenem(Drug information on imipenem)-cilastatin [Primaxin], or meropenem(Drug information on meropenem) [Merrem]) offer broad-spectrum activity against gram-negative bacteria (including P aeruginosa) and have therefore been assessed as single-agent therapies in patients with neutropenic fever.[15-20] These drugs offer an attractive alternative to penicillin- and aminoglycoside-containing regimens because of their wide spectrum of activity against gram-negative bacteria and their favorable toxicity profile.

For patients undergoing prolonged marrow aplasia, however, effective coverage of P aeruginosa and prevention of late-onset multidrug resistance are of paramount importance. Imipenem and cefepime(Drug information on cefepime) are, in addition, efficacious against some gram-positive organisms and anaerobes. The post-antibiotic effect associated with imipenem is particularly useful in the setting of profound neutropenia because it provides continued antibacterial activity in the absence of phagocytic cells.

Newer fluoroquinolones, including levofloxacin(Drug information on levofloxacin) (Levaquin), show moderate activity against anaerobes, as well as broad-spectrum activity against aerobic gram-negative and gram-positive bacteria. This, coupled with their ease of administration and favorable toxicity profiles, makes them attractive candidates for empiric treatment of febrile neutropenia. Unfortunately, bacteria that have acquired resistance to other fluoroquinolones may be less susceptible to these newer agents.

Gram-Positive Infections

Gram-positive infections have increased in frequency among patients with oncologic diseases, and now account for the majority of positive blood cultures in many institutions. Factors responsible for this increased incidence of gram-positive infections are listed in Table 2. These infections can be responsible for life-threatening illness in neutropenic patients, as exemplified by recent reports of overwhelming infections caused by viridans group strep-tococci (eg, Streptococcus mitis and Streptococcus sanguis) in patients treated with high-dose chemotherapy and/or radiation therapy,[21] and fatal outcomes in patients who received platelet transfusions contaminated with gram-positive bacteria.[22]

The dissemination of gram-positive organisms from possible sites of barrier breakdown can be suppressed by effective prophylactic therapy. Vancomycin(Drug information on vancomycin), a cell wall–acting glycopeptide antibiotic, is efficacious against a broad range of gram-positive organisms including many Staphylococcus isolates, Corynebacterium species, and other bacteria that are resistant to beta-lactam antibiotics.

In prospective clinical trials at centers where gram-positive infections are prevalent, vancomycin therapy begun empirically at the time of first infectious fever results in prompt resolution of fever, rapid clearance of local and/or disseminated gram-positive infections, and prevention of late-onset gram-positive infection.[23-24] In contrast, these benefits have not been realized in centers that have a lower prevalence of gram-positive infections.[25,26]

In the latter setting, vancomycin has been effective in treating and eradicating established infection in a timely fashion, whether it is added empirically for prolonged fever or selectively based on the suspicion or diagnosis of gram-positive infections (Table 3).[24,27,28] Prophylactic administration of vancomycin[29,30] or teicoplanin(Drug information on teicoplanin) (Targocid) to flush intravenous lines or for indwelling fluid locks has been shown to reduce the risk of catheter colonization and associated infections, although studies using systemic administration of either drug have yielded conflicting results.[31-34]

Vancomycin-Resistant Bacterial Infections—Since 1986, vancomycin-resistant enterococci (VRE) have emerged as an increasingly visible nosocomial problem, with epidemics reported recently at several hospitals. Risk factors for the acquisition of VRE include prolonged hospitalization, prior antibiotic exposure—particularly to cephalosporins and vancomycin—and admission to an intensive care unit or ward where the prevalence of VRE is high.

Isolates exhibiting the van A phenotype, including many strains of Enterococcus faecalis and Enterococcus faecium, have been studied most extensively at the level of genetic mechanisms. In these strains, an inducible plasmid-borne genetic transposition results in synthesis of cell wall peptide residues incapable of binding to peptidoglycan antibiotics; these strains are thereby able to avoid cell wall disruption by vancomycin or teicoplanin. The van B phenotype, also found in strains of E faecalis and E faecium, exhibits a somewhat lower level of resistance to vancomycin and is relatively susceptible to teicoplanin.

Enterococci that have either phenotype also tend to be intrinsically resistant to many other antibiotics, however, including beta-lactams (due to the presence of penicillin-binding proteins with decreased affinity), aminoglycosides, and quinolones.

Thus, therapeutic options for patients infected with these multiply resistant enterococci are limited. This problem has launched a search for newer antibiotic agents that might provide some activity against VRE. Unfortunately, in vitro activity against VRE isolates, identified in several classes of drugs (Table 4), has not readily translated into clinically effective therapy. Thus, the establishment and maintenance of vigorous infection control procedures remain the cornerstone of programs aimed at controlling this group of organisms, not only to limit their dissemination but also to prevent glycopeptide resistance in other, more virulent gram-positive pathogens, such as Staphylococcus aureus.

The recent descriptions of vancomycin-resistant S aureus arising in the setting of prolonged vancomycin treatment are a stark reminder of this threat.[35-37] A combination of careful and innovative antibiotic use and strict adherence to appropriate infection control procedures[38] are the only means for dealing with the expanding problems of antibiotic resistance in nosocomially transmitted organisms.

Clostridium difficile Infections—Infection control procedures are also important in limiting the spread of Clostridium difficile.[38] This gram-positive anaerobic bacillus is associated with colitis and diarrhea in patients with prolonged exposure to broad-spectrum antibiotics or some forms of chemotherapy, notably, cisplatin (Platinol) and high-dose cytarabine(Drug information on cytarabine).[39-41] In the patient rendered aplastic as a result of chemotherapy, infectious diarrhea presents special problems because of the limited utility of stool examination and the difficulty in distinguishing diarrhea due to infection from diarrhea due to chemotherapy or mucositis.

In addition, fevers associated with C difficile are apt to be mistaken for (or else obscured by) those associated with concurrent bacterial, viral, or fungal infection. In one study, the risk of VRE carriage was increased in patients who had previously contracted C difficile colitis.[42] Colitis also results in the hematogenous dissemination of enteric organisms, including VRE.[43]

Unfortunately, the discontinuation of antibiotics, clearly an important measure in the treatment of C difficile diarrhea, is infeasible in patients with febrile neutropenia. Recent research has focused on methods for the more rapid and accurate diagnosis of these infections, including tests based on the detection of bacterial toxin in stool samples, polymerase chain reaction (PCR)–based assays, and newer methods for increasing the yield from stool cultures. Anticlostridial antibodies are also being investigated for the prevention and treatment of this disease.[44]

Outpatient Antibiotic Therapy—Febrile, neutropenic patients are not all the same. The risk of adverse outcomes is related to the depth and duration of neutropenia, severity of mucositis, and/or the presence of indwelling vascular catheters. The presence of fungal colonization by multiple species with invasive potential, obvious sites of infection, positive blood cultures, or severe comorbidities also portend a poor outcome. Conversely, the risk of severe or life-threatening infections can be reduced by the prophylactic use of oral quinolones or granulocyte growth factors.

Limited studies have demonstrated the safety of using oral or once-daily intravenous antibiotics for the outpatient treatment of neutropenic fever in cancer patients.[45,46] However, there are insufficient data to recommend this approach for patients with acute leukemia, who generally experience profound aplasia and extensive treatment-induced oral and GI mucositis. Nevertheless, selected afebrile neutropenic patients with acute leukemia who have intact hematologic function (eg, during consolidation therapy) and who are receiving moderate rather than intensive chemotherapy may be managed as out-patients.[47]