Management of the Febrile Neutropenic Patient: A Consensus Conference

Infection in the immunocompromised host has been a cause for concern in the clinical setting and a topic of debate for decades. The balance in the struggle against this type of infection swings both ways. On one hand, there is an improved understanding of risk factors and a greater availability of diagnostic and treatment options. On the other, immunosuppressed hosts have increased in number, new risk factors—some which are yet to be identified—are emerging, the epidemiology of infections is constantly changing, and bacterial resistance is increasing. Hence, despite the efforts made, infection in the immunocompromised host still poses a serious problem that involves different medical specialties. In February 1998, a group of experts met in Miyazaki, Japan, and issued a set of recommendations for the use of antimicrobials in febrile neutropenia in Japan, based on the evidence available at the time [1]. Later on, the Japan Acute Leukemia Study Group reviewed the infectious complications in these patients [2]. Because >5 years has elapsed, it seems appropriate to assess now the validity of these recommendations, in the light of the experience acquired and the new evidence that has appeared. The impact of the Infectious Diseases Society of America (IDSA) guidelines on the management of these patients [3–7] should be underscored. Economic issues are also relevant, because cost-effectiveness is now considered a precondition of any health care intervention. Professor T. Masaoka, chairman of this meeting, discusses some of the issues reviewed during the conference related to a cost-effective approach to the management of these patients [8]. During the 1950s, the major etiologic agent among febrile neutropenic patients was Staphylococcus aureus. During the subsequent 2 decades, gram-negative bacteria emerged as the main group; infection with Pseudomonas aeruginosa was particularly common, and was associated with a high mortality rate. By the end of the 1980s and during the 1990s, gram-positive bacteria reemerged as the principal causative agents. The most common infective agent was now coagulase-negative staphylococci, although enterococci and S. aureus were also present [9–13]. The causes of this change in the epidemiology of the infection have not been clearly identified, but they may include more aggressive radiotherapy and chemotherapy, more profound and prolonged neutropenia, increased use of intravenous catheters, and selective pressure by antibiotics [10, 13]. Yet another trend towards change—namely, increasing involvement of gram-negative bacteria—seems to be occurring currently [14–16]. The changes in the etiology of bacteremia in febrile neutropenic patients have certainly been critical in the setting up of appropriate therapy. This issue is reviewed in depth in this supplement by Ramphal [17]. Furthermore, many “new” microorganisms have emerged. These were traditionally considered possible contaminants, but, during the past decade, they have become increasingly important and have been reported as pathogens in febrile neutropenic patients because of their potential to cause severe infections and their variable susceptibility to standard antimicrobials [9]. With respect to gram-positive microorganisms, an increase in the number of infections due to coagulase-negative staphylococci has been observed worldwide, so much so that these organisms are recognized to be the most common pathogens causing nosocomial bacteremia in the general hospital population and, in particular, in the neutropenic host [18]. Coagulase-negative staphylococci are one of the main components of the skin flora and for years have been considered a contaminant. The patient's own flora is currently considered the main source of infection, although even the air of hospital units for patients with neutropenia has been reported to be a potential cause of outbreaks [19]. Infections with coagulase-negative staphylococci are quite indolent and are associated with a low mortality [20, 21]. Viridans group streptococci belong to the standard flora of the oropharynx and gastrointestinal tract; foci of infection probably originate from these sites. Although viridans streptococci were originally associated with infectious endocarditis, they have become one of the main causes of bacteremia among neutropenic patients and, unlike coagulase-negative staphylococci, are associated with high morbidity and mortality [20]. Enterococcus species are increasingly identified, particularly as causes of nosocomial bacteremia. The most noteworthy aspect is the change in the distribution of the species: the number of infections with Enterococcus faecium is increasing more rapidly than are infections with Enterococcus faecalis [12]. The “new” gram-positive pathogens include cocci such as Stomatococcus mucilaginosus and Leuconostoc species, which cause catheter-related bacteremia [9, 10]. Similarly to the latter, Pediococcus species are isolated from feces, wounds, urine, and blood [12]. Among the gram-positive rods, Corynebacterium jeikeium has been associated with endocarditis, catheter-related bacteremia, cutaneous lesions, and nodular pulmonary infiltrates [10]. Lactobacillus species cause bacteremia, endocarditis, meningitis, intra-abdominal abscesses, and pneumonia and is associated with significant mortality, as high as 45% [9, 10]. Rhodococcus equi has been implicated in suppurative pneumonia with pulmonary abscesses and empyema [12]. Clostridium septicum is an anaerobe that produces a toxin that induces sepsis and metastatic myonecrosis. It is the most frequent cause of typhlitis in the neutropenic patient [10]. Gram-negative microorganisms, represented mainly by P. aeruginosa and Enterobacteriaceae such as Escherichia coli, Klebsiella species, and Enterobacter species are still a problem, despite a decline in their incidence over the last few years, because of their traditional association with high rates of mortality and the alarming appearance of multidrug-resistant strains [10]. A study of current antimicrobial usage in Japan shows that P. aeruginosa continues to be the most common pathogen (accounting for 23.5% of isolates) (M. Yoshida and R. Ohno, personal communication). The “new” gram-negative organisms include Stenotrophomonas maltophilia, involved mainly in nosocomial bacteremia. Alcaligenes xylosoxidans and Burkholderia cepacia are related to catheter-associated sepsis. Capnocytophaga species have been associated with bacteremia in bone marrow transplant recipients, with severe involvement of the oral mucosa. Among the anaerobes in a strict sense, Fusobacterium nucleatum causes bacteremia, ulcerative pharyngitis, and nodular pulmonary infiltrates by septic emboli. Leptotrichia buccalis causes bacteremia with extensive mucosal involvement in severely immunosuppressed patients [10, 12]. Mycobacteria are not a usual cause of infections in patients with neutropenia. The most typical are those that grow rapidly, especially Mycobacterium chelonei and Mycobacterium fortuitum, implicated mainly in catheter-related infections [22]. Nevertheless, the increase in the number of cases of tuberculosis among the HIV-infected population suggests that the prevalence of infections with mycobacteria among patients with neutropenia may increase in the future [23]. As concerns fungi, mycoses are among the most severe infections in febrile neutropenic patients, not only because their incidence has increased [23], but also because of the new illnesses produced by traditional fungi and the appearance of new pathogens [24]. An additional problem is the difficulty in clinical recognition, the scarcity of diagnostic tools for identification, and the high incidence of associated mortality, probably related to delayed diagnoses [25]. Candida species cause infections ranging from superficial lesions to organ involvement and severe systemic candidiasis, which in the patient with neutropenia usually has a nonspecific presentation, with little endophthalmitis [25]. Recently, severe chronic infections have been reported, including hepatosplenic candidiasis that usually develops after the patient has recovered from neutropenia [26, 27]. Aspergillus species produce mainly invasive pulmonary aspergillosis that can recur after its clinical resolution if fungus balls persist in the pulmonary cavities. It can even cause severe pulmonary exsanguinating hemorrhage [23, 24] or can disseminate [24, 28]. The new illnesses described in febrile neutropenic patients include primary cutaneous aspergillosis associated with the use of intravenous catheters, which can disseminate to underlying organs [24]. The fungi now include “new” pathogens [28]. Fusarium species [25], which causes disseminated infection with involvement of noble organs and invasion of blood vessels leading to thrombosis and infarction that closely resemble Aspergillus infection. The clinical presentation is extremely severe and difficult to treat [9]. Trichosporon beigelii and Blastoschizomyces capitatus can cause disseminated infection with organ involvement and cutaneous lesions that progress from nodules to ulcers, with a high mortality rate [25]. Scedosporium apiospermum and Scedosporium prolificans can produce esophagitis, sinusitis, pneumonitis, osteomyelitis, meningitis, and disseminated infection [24, 25]. Finally, new species of classic fungi are also emerging, such as Candida krusei, Candida lusitaniae, Candida utilis, Candida dubliniensis, and Candida guillermondii. The impact of fungal infection and its empirical treatment in febrile neutropenic patients is reviewed by Wingard [29] in this supplement. Viral infections are frequent in patients with leukemia and lymphoma, but they are associated more with cellular immunosuppression than with neutropenia. Primary infections, reinfections, or reactivations can occur. The principal viral agents of infection in the patient with neutropenia are the herpesvirus family, particularly cytomegalovirus, which can produce hepatitis, pneumonia, and involvement of the gastrointestinal tract [23, 30], with high rates of morbidity and mortality. Unlike the situation in the patient with AIDS, retinitis is uncommon. Viruses also include the “new” pathogens, such as the human metapneumonia virus, identified in 2001, although it may have been causing infection since the 1950s in the form of respiratory involvement with bronchiolitis and pneumonia very similar to those produced by respiratory syncytial virus [31, 32]. Similarly, parasite infestations are more related to cellular immunosuppression in these patients. Toxoplasma gondii can cause cerebral abscesses, Cryptosporidium species can cause enteropathy, and Strongyloides stercoralis can cause diarrhea [30] and associated bacteremia secondary to Strongyloides infestation, in immunocompromised patients [33–35]. Of concern among gram-positive bacteria is resistance to glycopeptides, mainly coagulase-negative staphylococci, which are most resistant to teicoplanin [36, 37], and Enterococcus, which shows its greatest resistance to vancomycin. This represents a serious threat because, in vitro, vancomycin resistance genes have been transferred to strains of S. aureus and coagulase-negative staphylococci, and there is concern that this may also happen widely in vivo [38]. In fact, 2 cases of infection caused by vancomycin-resistant S. aureus containing the vanA gene have been described already; in one of them, a vancomycin-resistant strain of E. faecalis was isolated from the same swab specimen [39, 40]. The occurrence of outbreaks of infection due to Enterococcus species has been connected with sources in the hospital environment, and infections are usually transmitted from patient to patient via the hands of health care workers [11]. Resistance to β-lactams can be mediated by β-lactamase production, a mechanism that extended rapidly among Staphylococcus species after the introduction of penicillin [41, 42] but had not been described in Enterococcus species until 1983 [43]. β-lactamase production in Enterococcus species is associated almost universally with resistance to aminoglycosides and results in a highly pathogenic strain [44]. The other mechanism of resistance to β-lactams is a defective affinity of penicillin-binding proteins, responsible in Staphylococcus species for the recent pandemic of infections by methicillin-resistant Staphylococcus strains [45]. The appearance and rapid spread of strains of Enterococcus species with high-level resistance to aminoglycosides has aroused concern, because they can result in more-severe disease with a worse prognosis [44–47]. α-Hemolytic streptococci are acquiring a high degree of resistance [48] and cause disease with a severe clinical presentation and with high mortality among febrile neutropenic patients [49]. Of note among gram-negative microorganisms is quinolone resistance in E. coli [50]. Its appearance has been associated with an increased use of fluoroquinolones as prophylaxis for febrile neutropenia which, in only 10 years, rose from 1.4% of prescriptions in 1983 to 45% of prescriptions in 1993 [10]. These resistant strains are not interrelated [51], indicating their diversity and the fact that a single clone cannot explain their epidemiology. An alarming increase has been reported in the incidence of multidrug-resistant Enterobacter strains; this can be attributed to the induction of chromosomal β-lactamases by the frequent use of third-generation cephalosporins [52]. P. aeruginosa has shown increased resistance to imipenem, cephalosporins, and ciprofloxacin. S. maltophilia is often susceptible only to cotrimoxazole, as are Alcaligenes species [52, 53]. Acinetobacter species are becoming a threat as nosocomial pathogens because they tend to be resistant to multiple antimicrobials; the carbapenems are the agents most active against them [54, 55]. Special mention should be made of the appearance among gram-negative organisms of extended-spectrum β-lactamases, which confer resistance to all of the β-lactams except the carbapenems. First reported in 1983, the plasmid character of the genes responsible conditions a rapid and high capacity for dissemination, thus constituting a severe nosocomial problem. The main risk factor for the appearance of extended-spectrum β-lactamase–producing microorganisms is antibiotic pressure. Extended-spectrum β-lactamase–producing potential has been observed mostly in Enterobacteriaceae, but it can occur also in other gram-negative organisms, such as P. aeruginosa. Extended-spectrum β-lactamase–producing bacteria are mostly multidrug resistant; they frequently exhibit a lack of susceptibility to cotrimoxazole, aminoglycosides, quinolones, and tetracyclines [48, 56–58]. In the majority of fungi, drug resistance has been related to the excessive or inappropriate use of antifungals, mainly for prophylaxis, which can generate more resistant pathogens [59] or lead to the emergence of new pathogens with diminished susceptibility. Infections with Candida species are increasingly difficult to treat. Species such as C. lusitaniae, C. guillermondii, and Candida lipolytica are more prone to be resistant to amphotericin B; C. krusei shows inherent resistance to fluconazole, Candida glabrata is also frequently resistant to fluconazole [60], and 8%–10% of species show a lack of susceptibility to 5-fluocytosine. Other fungi, such as Fusarium species, Scedosporium species, and T. beigelii, are usually resistant to amphotericin B, even at high doses [25]. Local data on the epidemiology of these infections, the incidence of certain pathogens, and resistance patterns are of the utmost importance in the determination of options for empirical treatment. Kanamaru and Tatsumi [61] present an observational study of the microorganisms involved and their relation to outcome. The neutropenic patient is a complex host in whom several factors come together to generate the patient's proneness to infection. When the different defense mechanisms, which usually act synergistically, are altered, the patient becomes more vulnerable to a wide range of infections. A recent approach with a significant impact on the management of these patients is assessment of risk of developing serious complications. Indeed, there is a fairly precise understanding of the factors implicated in the evolution of these patients' conditions, and we can (and must) prevent their conditions from deteriorating. Stratification of patients into low-risk and high-risk categories has paved the way for efficient treatment strategies [13, 62–68]. Klastersky [69] reviews the state of the art in risk assessment. The management of these patients is critical, and there is a large body of evidence documenting significant differences among different treatment regimens. It is therefore mandatory to use diagnostic and therapeutic resources in the most appropriate way. To achieve this, scientific institutions have issued different consensus documents that, in the light of the available evidence, propose quite specific courses of action for febrile neutropenic patients [3, 6, 7, 68, 70–73], and there has been much emphasis on the need to agree on some important issues in these documents; for example, to provide a single definition of treatment response [74]. The application of these guidelines has led to a greater use of therapeutic resources [75]. Also, prophylaxis against the infections in these highly susceptible patients has largely been studied, and Yoshida and Ohno [76] review the currently available information on this subject. Most infectious processes in the patient with neutropenia evolve with minimal clinical expression. Hence, the probability of infection should be considered when an otherwise unexplained change in the patient's clinical status takes place [77]. An accurate microbiological diagnosis plays an essential role in the clinical management of these patients [78–80]. Unfortunately, microbiological testing is often limited to obtaining blood and urine samples for culture [81]. Failing to obtain the appropriate specimens for culture before initiating empirical antibiotic treatment whenever an infection is suspected can delay the true diagnosis and recognition of a nonbacterial infection and lead to the unnecessary use of broad-spectrum antibiotic therapy, thereby increasing the risk for adverse events and superinfections with resistant microorganisms. It can also hinder the determination of whether the infection was present initially after the therapeutic failure, because the previous use of antimicrobial drugs can lead to negative culture results, as Masaoka [8] rightly pointed out at this meeting. Rolston [82] addresses the most relevant concepts in the 2002 IDSA guidelines for the use of antimicrobial agents in treating patients with neutropenia with cancer. Yoshida and Ohno [83] describe the results of a survey on the current management of infectious patients with acute leukemia in Japan and the documents drawn up during the Second Consensus Conference held in Honolulu on 8 and 9 March 2003. The availability of new, wide-spectrum antimicrobials has provided adequate coverage for recently emerged pathogens and has facilitated the use of monotherapy, used on the basis of risk categorization, as discussed by Ohyashiki [84]. Also, Tamura and a group of distinguished colleagues [85] describe a randomized trial comparing cefepime as monotherapy or in combination with amikacin for the empirical treatment of febrile neutropenia. The final recommendations issued by the group of experts gathered in this meeting are presented as an algorithm. Urabe [86] provides the definitions on which the recommendations are based, whereas Tamura [87] and Masaoka [88] describe the chain of options that make up the algorithm. This consensus document resulted from a wide range of evidence, which was analyzed in depth and discussed by the experts. We trust that, once again, it will be of help for the appropriate management of febrile neutropenic patients.