Infections remain an important cause of morbidity and mortality in cancer patients, mainly those affected by haematological malignancies and those undergoing stem cell transplant (SCT).1–3 With the improvement of cancer care in the past decades, patients live longer, and immunosuppression from the underlying disease and from more intensive and prolonged treatments renders more and more of them susceptible to infections.
Infections remain an important cause of morbidity and mortality in cancer patients, mainly those affected by haematological malignancies and those undergoing stem cell transplant (SCT).1–3 With the improvement of cancer care in the past decades, patients live longer, and immunosuppression from the underlying disease and from more intensive and prolonged treatments renders more and more of them susceptible to infections.
The epidemiological evolution of infectious complications in cancer patients is characterised not only by the increasing incidence of some opportunistic infections, such as invasive fungal diseases (IFDs), but also by the emergence of antimicrobial resistance in several bacteria and some fungi. Additionally, viruses are a serious and frequently underestimated cause of infection in cancer patients, particularly those subjected to SCT procedures, in whom careful surveillance and an antiviral prophylaxis strategy are required.
Thanks to the availability of new laboratory and imaging diagnostic tools and to the development of tailored surveillance strategies, a better knowledge of the epidemiology and risk/prognostic factors of infections has been gained in cancer centres in recent years. This epidemiological awareness is crucial to define prevention strategies adapted to the various categories of patients with different infection risks, and to chose treatment protocols likely to contain the above-mentioned phenomenon of resistance to antimicrobials. In this article, recent epidemiological findings on, and current prevention strategies against, bacterial, fungal and viral infections in cancer patients are reviewed and critically commented on.
Epidemiology of Infections in Cancer Patients
Despite increasing attention to the clinical and therapeutic aspects of infection in cancer patients, few data are available on the incidence, microbiological characteristics and clinical outcomes of infectious complications in this population. Recent epidemiological data almost exclusively derive from retrospective studies and few prospective data are available. The susceptibility pattern to bacterial and fungal pathogens has been occasionally evaluated in the cancer population and most data derive from large microbiological studies in patients with various underlying conditions. Again, most data in the cancer setting derive from neutropenic patients affected by haematological malignancies and/or undergoing SCT.
Bacterial Infections
Over the past decades, the epidemiological spectrum of bloodstream bacterial isolates obtained from febrile neutropenic patients has continuously changed. During the 1960s and 1970s, gram-negative bacilli represented the predominant pathogens causing life-threatening infections. During the 1980s and 1990s, gram-positive organisms became more common, probably due to the increased use of venous catheters and systemic prophylaxis with fluoroquinolones, which can allow entry of, and colonisation by, gram-positive skin and intestinal flora.4–7 Currently, coagulase-negative staphylococci continue to be the most common blood isolates; however, the emergence of infections by Enterobacteriaceae (e.g., Escherichia coli and Klebsiella species) and non-fermenting gram-negative pathogens (e.g., Pseudomonas aeruginosa and Stenotrophomonas maltophilia) has been observed in several centres.8–19
In particular, drug-resistant gram-negative bacteria species are causing an increasing number of infections in febrile neutropenic patients. Klebsiella species and E. coli strains with acquired extended-spectrum beta-lactamase (ESBL) genes frequently show a broad range of beta-lactam antibiotic resistance. These ESBL producers are often only susceptible to carbapenems, antibiotics with a very broad spectrum of activity but whose use may induce the selection of bacteria with even greater resistance patterns. Indeed, the use of carbapenems increases the risk of infections by carbapenemase-producing Klebsiella species and P. aeruginosa strains that are resistant to carbapenems and other classes of antibiotics. This phenomenon may represent a devastating problem, with the spread of infections by multi-drug-resistant (MDR) gram-negative strains, considering also that the molecules recently added to the antibacterial armamentarium are mainly oriented against gram-positive bacteria.
Information regarding incidence, risk factors and clinical findings for resistant gram-negative bacteraemia in cancer patients and its outcomes is scarce and dispersed. Only a few single-centre experiences have been published in recent years and the real epidemiological impact of such infections is little known. In a prospective survey in an adult Spanish cancer centre conducted during the period 2006–2009, about 50 % of bacteraemias were caused by gram-negative bacilli and 14 % of them were caused by MDR strains. The most frequent mechanism of resistance was ESBL production (45 %), mainly in E. coli, followed by AmpC cephalosporinase hyperproduction (24 %). Patients with MDR gram-negative bacteraemias more frequently received inadequate initial antibiotic therapy (69 % versus 9 %, p<0.001), and time to adequate therapy (after 48 hours) was longer in this group (41 % versus 4 %, p<0.001). Patients in the resistant group more frequently required intensive care unit admission (14 % versus 5 %, p=0.023), had greater need of mechanical ventilation (14 % versus 3 %, p=0.005) and had a higher overall case fatality rate (41 % versus 21 %, p=0.003).15
A very high rate of resistant gram-negative isolates in patients with haematological malignancies was observed in an Italian centre.18 Out of 62 episodes of bacteraemia caused by E. coli, overall incidences of ESBL production and fluoroquinolone resistance were 41.9 % and 62.9 %, respectively. The overall 30-day mortality rate was 21 %. In a multivariate analysis, the significant predictors of mortality were inadequate initial antimicrobial therapy (odds ratio [OR]=14.96, 95 % confidence interval [CI] 1.95–114.51, p=0.009), infection caused by ESBL-producing isolates (OR=8.84, 95 % CI 1.48–52.91, p=0.01) and prolonged neutropenia (OR=8.10, 95 % CI 1.29–50.57, p=0.02).
A prospective study of infections in nine Italian tertiary care centres or university hospitals among adult patients diagnosed with haematological malignancies was started in 2009.19 The analysis of patients with P. aeruginosa bacteraemias showed 71 % MDR strains. In particular, the percentages of resistance to carbapenems (imipenem and meropenem), anti-pseudomonal cephalosporins (ceftazidime and cefepime), amikacin and ciprofloxacin were 60 %, 42 %, 50 % and 66 %, respectively, whereas the percentage of resistance to piperacillin was 24 %. Mortality within 30 days of the first positive blood culture occurred in 40 % of MDR P. aeruginosa bacteraemias and in 9 % of non-MDR P. aeruginosa bacteraemias. In a multivariate analysis, inadequate initial antibacterial therapy was independently associated with mortality (p=0.006).
All these studies show the increasing epidemiological burden of severe infections by resistant bacterial pathogens. A crucial problem in the management of such infections is the choice of prophylaxis and initial antibacterial therapy, considering that it may dramatically affect the outcomes of such complications. A continuous microbiological survey at each cancer centre is required to monitor the trend of resistant micro-organisms and to guide the use of antimicrobials.
Fungal Infections
IFDs, especially those caused by Candida and Aspergillus species, historically represent a major cause of death in cancer patients, especially after intensive chemotherapy and receipt of SCT. However, a continuous modification in epidemiology and outcomes has been observed over the last decades, as a result of prevention strategies and changes in host characteristics and treatment modalities.20–22 Another factor that may greatly impact on epidemiological findings in IFDs is the type and quality of surveillance studies. In fact, different results have been obtained according to whether studies were retrospective or prospective, and single-centre or multicentre. Until a few years ago, most epidemiological and outcome data regarding IFDs in cancer patients came from single centres using different analysis criteria, which raised questions about the ability to generalise their findings. Only in the last few years have multicentre surveillance studies, which also evaluated the impact of host and chemotherapy and/or transplant factors on outcomes, been published.23–25
Very recent clinical data from a general population of patients with candidaemia were extracted from the Prospective Antifungal Therapy Alliance database, a comprehensive North American registry that collects information regarding IFDs.23,24 Contemporary epidemiology and outcomes of candidaemia in multiple centres were evaluated in a total of 2,019 patients enrolled from 2004 to 2008.23 Overall, patients treated in general medicine (66 %), who underwent non-transplant surgery (33 %) or had solid tumours (17 %) were the most represented, whereas 9.8 % and 2.9 % of the candidaemia episodes were documented in patients with haematological malignancies and in SCT recipients, respectively. Non-albicans Candida species accounted for the large majority of isolates (72.6 % in haematological malignancies, 77.6 % in SCT and 52.4 % in solid tumours).
The same registry was evaluated to analyse IFDs that occurred specifically among SCT recipients.24 Out of 250 IFDs, invasive aspergillosis was the most frequent (59 %), followed by candidiasis (25 %) and infections caused by other moulds (14 %). A surprising finding was the unexpected lower 12-week mortality in patients with invasive aspergillosis compared with patients with invasive candidiasis (36 % versus 49 %, respectively). However, improved survival after the diagnosis of invasive aspergillosis was consistent with that observed in prior single-centre and multicentre studies in the US and Europe.26–29 The better knowledge of epidemiology as well as of risk and prognostic factors, and the recent advances made in the prevention, diagnosis and treatment of Aspergillus infections probably justify such advances in the management of IFDs.
For fungi also does the reduced susceptibility to antifungal drugs represent a challenging problem. The phenomenon of Candida species resistance to triazoles has been largely reported for fluconazole and itraconazole since their first use more than two decades ago, and it continues to be an important issue for the second-generation triazoles posaconazole and voriconazole.30–34 Resistance to triazoles has also been observed for Aspergillus species in some European countries, although, according to multinational surveys, the phenomenon seems uncommon. An in vitro study of triazole cross-resistance performed among more than 700 clinical isolates of Aspergillus species collected between 2000 and 2006 as part of a global antifungal surveillance programme showed a pattern of resistance (minimum inhibitory concentration [MIC]>2 μg/ml) for itraconazole in about 2 % of isolates, and in less than 1 % of the isolates for voriconazole and posaconazole.35
Viral Infections
Several viruses may cause infections in patients with cancer, and important advances have been made in recent years in their molecular diagnosis. Herpesviruses and hepatitis viruses may cause serious infections after chemotherapy or after SCT, generally as a result of viral endogenous reactivation. Up to 80 % of adult cancer patients are herpes simplex virus (HSV)-seropositive and HSV lesions occur in more than 50 % of these patients as a consequence of the reactivation of a latent virus – whereas primary infection is unusual. Among seropositive recipients without prophylaxis, following allogeneic SCT, the risk of HSV infection is approximately 80 %, the risk of cytomegalovirus (CMV) infection is 20–30 %, and the risk of varicella zoster virus (VZV) infection is 20–50 %.36–39 Reactivation of chronic hepatitis B in patients undergoing immunosuppressive or antineoplastic treatment occurs in approximately 20–50 % of hepatitis B surface antigen (HBsAg)-positive patients and can result in fulminant hepatitis.40 Patients with malignant lymphoma, especially those treated with anthracyclin-containing chemotherapy, are at significantly higher risk.41 HBsAg-positive allogeneic SCT patients are at risk of developing severe and possibly fatal hepatic disease, and therefore anti-hepatitis B surface (HBs)-negative donors for HBsAg-positive patients should be vaccinated before stem cell collection. Conversely, transplantation of an HBsAg-negative patient with stem cells from an HBsAg-positive donor is associated with a high risk of transmission, but few patients develop aggressive acute infection or chronic hepatitis B.42–44
The detection and diagnosis of infections associated with community respiratory virus have improved substantially with the use of various polymerase chain reaction (PCR) techniques, but the routine application of these tests is limited to a few settings in cancer patients; therefore, the real epidemiological impact of respiratory virus infections is probably underestimated. Early identification using rapid and reliable diagnostic methods and strategies to contain the infection and prevent nosocomial spread are essential to prevent the devastating consequences caused by outbreaks of community respiratory viruses.
Prophylaxis Against Infections in Cancer and Stem Cell Transplant Patients
Until recently, the impact of prophylaxis against infections in cancer patients had been controversial, but new data now suggests that chemoprophylaxis might be able not only to reduce the incidence of infectious complications, but also to allow the continuation of treatment of the underlying malignancy, with a possible impact on long-term overall survival. Current European and US guidelines on antibacterial, antifungal and antiviral prophylaxis have been summarised in Table 1 and Table 2, respectively.45–49
Antibacterial Prophylaxis
The risk of bacterial infection is directly related to the severity and duration of neutropenia. The practice of chemoprophylaxis for the prevention of bacterial infections in neutropenic cancer patients derived from the observation that most of the bacterial pathogens causing infection came from the patient’s endogenous flora of the gastrointestinal tract.1 Oral and gastrointestinal chemotherapy-induced mucositis represents a major risk of bacterial entry directly into the bloodstream. Selective intestinal decontamination with oral fluoroquinolones, which not only operate directly in the gastrointestinal tract but also act systemically, has been the most attractive antibacterial prophylactic strategy in neutropenic patients for 20 years. Patients who derive particular benefit from prophylaxis are those with a duration of neutropenia of more than seven days, as they are at high risk of bacterial infections.
In view of the need to re-evaluate the role of prophylaxis with fluoroquinolones in neutropenic cancer patients after a wide use in clinical practice, and to better define the categories of patients who might most benefit from it, two large multicentre controlled trials were conducted in Europe and the results published in 2005.50,51 The Gruppo Italiano Malattie Ematologiche dell’Adulto (Italian adult haematological disease group) (GIMEMA) conducted a double-blind, placebo-controlled trial in 760 hospitalised high-risk adult patients with prolonged neutropenia (<1,000 neutrophils/mm3 for more than seven days).50 The trial included patients with acute leukaemia, lymphoma and solid tumours receiving high-dose chemotherapy. Participants were randomised to receive levofloxacin or placebo from the start of chemotherapy until resolution of neutropenia. The study demonstrated a statistically significant reduction in the incidence of fever in patients receiving levofloxacin compared with the placebo group (65 % versus 85 %, p<0.001), associated with a non-significant decrease in mortality in the levofloxacin arm.
The second study, the British Simple investigation in neutropenic individuals of the frequency of infection after chemotherapy ± antibiotic in a number of tumours (SIGNIFICANT) trial, look ated 1,565 patients receiving cyclical chemotherapy for solid tumours or lymphoma at risk of temporary neutropenia and classified as being at low risk.51 Patients randomly received either levofloxacin or placebo for seven days during the expected neutropenia period. A significant reduction in febrile episodes was documented in the levofloxacin group during the first cycle of chemotherapy (3.5 % versus 7.9 %, p<0.001) and all cycles of treatment (10.8 % versus 15.2 %, p=0.01), but no effect on documented infections was observed.
These two studies indicate a potentially important role of levofloxacin prophylaxis in high-risk patients with prolonged neutropenia, but not in low-risk patient populations.
The major drawback of the routine use of prophylaxis is the emergence of fluoroquinolone resistance in both gram-positive and gram-negative organisms. The phenomenon has been observed since the 1980s and the clinical implications of fluoroquinolone resistance continue to be investigated.52–57 A systematic review of the effect of quinolone prophylaxis in neutropenic patients showed that prophylaxis was associated with a decreased overall incidence of infection, a non-significant increase in colonisation with resistant organisms and no difference in the number of infections caused by resistant organisms.57
In summary, the problem of resistance during fluoroquinolone prophylaxis in cancer patients and its clinical implications remain controversial. In any case, fluoroquinolone resistance with possible cross-resistance to other antibiotics should be carefully monitored and the efficacy of the practice of antibacterial prophylaxis periodically re-evaluated.
Antifungal Prophylaxis
Antifungal prevention strategies are based on environmental precautions and antimicrobial treatment. While there is general agreement on the role of air filtration in the control of airborne filamentous fungal infections, the indication for pharmacological prophylaxis is still debated.58–60 The use of antifungals for prophylaxis against deep-seated IFDs has become standard practice of care in some haematological and transplant patients, whereas, in patients with solid tumours, who are at lower risk of invasive infections, antifungal prophylaxis is mainly directed to the prevention of superficial mycoses.
Until a few years ago, fluconazole and, to a lesser extent, itraconazole were the only drugs recommended for primary prophylaxis against Candida infection in neutropenic patients and allogeneic SCT recipients.61,62 However, a major limitation of a Candida-oriented prophylaxis is its lack of activity against moulds, which now represent the most frequent cause of IFDs in these patient populations. In recent years, awareness of the epidemiological impact of invasive aspergillosis and of infections caused by less common moulds, including zygomycetes, Fusarium species and Scedosporium species, has increased worldwide. At the same time, new broad-spectrum and well-tolerated antifungal drugs, in particular second-generation triazoles and echinocandins, have become available, and prospective controlled trials have been conducted to investigate their ability to prevent IFDs in high-risk haematological patients.63–69
Posaconazole and voriconazole – second-generation triazoles with a broad spectrum of activity – have been compared with fluconazole or itraconazole in phase III clinical trials of prophylaxis in leukaemia and SCT patients.63–66 Posaconazole prophylaxis was associated with a significantly lower rate of invasive aspergillosis, both in neutropenic patients with acute myeloid leukaemia and in allogeneic SCT recipients with graft-versus-host disease (GVHD), but only in the leukaemia setting did patients who received posaconazole prophylaxis experience a significantly longer survival.63,64
Two controlled studies of primary antifungal prophylaxis with voriconazole have been conducted in allogeneic SCT patients.65,66 In the first study, in which voriconazole was compared with fluconazole, only a trend towards fewer Aspergillus infections was observed with voriconazole, and the fungal-free survival (primary endpoint of the study) was similar in the two groups.65 However, it should be considered that, in the subgroup of patients with acute myeloid leukaemia – a population at higher risk of infection – voriconazole significantly reduced IFDs and improved fungal-free survival compared with fluconazole.70 In the second study, voriconazole was superior to itraconazole considering a primary composite endpoint that included efficacy and tolerability criteria; however, no difference in the incidence of IFDs was observed in the two arms, probably due to the overall low rate of infections observed in the study.66
The clinical use of both posaconazole and voriconazole requires knowledge of the pharmacokinetic characteristics of these drugs that explain why their concentration in blood varies between patients and in an individual patient over the course of treatment. Voriconazole metabolism is particularly affected by the different ability of patients to metabolise the drug via the CYP2C19 P450 enzyme, whereas erratic gastrointestinal absorption represents a frequent problem in patients receiving posaconazole. For both antifungals, the metabolism may be affected by the pharmacological interaction with other drugs. In view of the above, therapeutic drug monitoring is strongly suggested during voriconazole and posaconazole treatment to monitor whether target through blood levels are achieved.71–75
Several open issues regarding prophylaxis against IFDs in patients with haematological disorders deserve careful consideration. The recommendations contained in the international guidelines reflect the important progress made in the prevention of IFDs, including those caused by filamentous fungi; however, no consensus has been reached regarding the optimal prophylaxis against IFDs in the complex scenario of haematological disorders, particularly in the transplant setting. This problem was recently underlined in a consensus process conducted by the Gruppo Italiano Trapianto di Midollo Osseo (Italian bone marrow transplant group) (GITMO), which observed that the recommendations from international guidelines entail various problems, such as the lack of an approved mould-active prophylaxis during the engraftment phase in allogeneic SCT and the lack of an intravenous (IV) formulation of posaconazole, which limits the use of this drug in patients unable to tolerate oral medications or with altered intestinal absorption.76
Additional and well-designed studies are needed, not only to evaluate the efficacy of new antifungal drugs, but also to define risk stratification criteria and tailored prevention strategies. Being aware of, and keeping up-to-date with, new epidemiological data is required to individualise specific measures aimed at the prophylaxis against IFDs in the different clinical settings encountered in haematological disorders.
Antiviral Prophylaxis
Primary HSV infection in cancer patients, including those affected by leukaemia, is unusual, and antiviral drug prophylaxis is thus not recommended in HSV-seronegative patients during chemotherapy or after SCT. Conversely, prophylaxis with an HSV-active agent, such as aciclovir, should be offered to all HSV-seropositive autologous or allogeneic SCT recipients and patients with acute leukaemia undergoing induction or reinduction therapy.47,49
Seronegative leukaemic patients and SCT recipients are at high risk of varicella after face-to-face contact of five minutes or more with a person who has varicella or intimate contact (touching or hugging) with a person who has herpes zoster. Patients residing in the same household as a contagious person, or staying in hospital in the same room as, or (in large wards) in a bed adjacent to, a contagious person, are also at risk. Passive immunisation with IV varicella zoster-specific immunoglobulins or IV normal immunoglobulin should be given as soon as possible after exposure (<96 hours) to VZV-seronegative cancer patients on intensive chemotherapy, receiving steroids or undergoing SCT with chronic GVHD. The efficacy of antiviral agents for post-exposure prophylaxis in leukaemic patients and recipients of SCT is uncertain, but uncontrolled studies seem to suggest that aciclovir, valaciclovir or famciclovir prophylaxis may reduce the incidence of varicella and its severity. Although most SCT recipients are VZV-seropositive, they are at risk of virus reactivation for a prolonged period after transplant. Therefore, for VZV-seropositive allogeneic SCT recipients, prophylaxis with oral aciclovir or valaciclovir is recommended for one year, or longer in the presence of GVHD and immunosuppressive therapy.47
The different antiviral strategies for CMV disease include the use of chemoprophylaxis, pre-emptive therapy or treatment of symptomatic infection. A pre-emptive antifungal strategy based on the monitoring of CMV infection by antigenaemia and/or quantitative PCR represents the most widely used approach, not only in the allogeneic SCT patient, but also in other categories at risk of CMV infection and disease, such as patients with chronic lymphocytic leukaemia undergoing alemtuzumab therapy. The frequency of CMV infection in other subgroups of cancer patients may be high, but the risk of evolution to CMV disease is very low; therefore routine surveillance is unnecessary and prophylaxis not recommended in these patients.48 As an alternative to pre-emptive therapy, IV ganciclovir prophylaxis is an effective strategy for the prevention of CMV disease in subgroups of allogeneic SCT patients at high risk of CMV disease, but acyclovir or valaciclovir at high doses can also be used. In any case, their use must be combined with monitoring and a pre-emptive approach.
Reactivation of hepatitis B is a well-characterised syndrome associated with the reappearance or rise of hepatitis B virus (HBV) DNA in the serum of a patient with previously inactive or resolved HBV infection, and is frequently accompanied by the reappearance of early or late hepatic disease activity. Reactivation of HBV has been reported not only in HBsAg-positive patients undergoing systemic chemotherapy, but also in a proportion of HBsAg-negative patients with hepatitis B core (HBc) antibody and/or HBs antibody. HBsAg-positive patients with haematological malignancies undergoing immunochemotherapy are at risk of developing severe hepatic disease. However, the most dramatic examples of HBV reactivation have been described in patients undergoing allogeneic SCT; in fact, vaccination before stem cell collection has been recommended for anti-HBs-negative donors to HbsAg-positive patients. Guidelines published by the US Centers for Disease Control and Prevention recommend that all patients about to receive chemotherapy for malignant disease be tested for HBsAg before cancer treatment is initiated.77 Controlled clinical trials and several subsequent meta-analyses have shown that prophylaxis with nucleoside analogues (e.g., lamivudine) decreases the incidence of HBV reactivation and the frequency of clinical hepatitis and death from HBV-associated liver injury in patients undergoing chemotherapy or SCT.78–84 Such therapy should be commenced before the start of chemotherapy or before transplant and should be continued for at least six months after the end of chemotherapy, or for longer in case of long-term immune suppression.85,86
Conclusions
Infectious complications in cancer patients continue to represent a challenging issue that greatly affects the overall care of the underling malignancy. Knowledge of the epidemiological phenomena of infections in such immunocompromised patients is required to better define surveillance and prevention strategies. Most epidemiological studies have been conducted in large patient populations with different tumour types, treatments and other patient characteristics, and may be unable to provide specific information for individual patient groups. Prophylaxis studies too have been conducted in large populations with some common risk factors but with a variety of different clinical characteristics. Consequently, current guidelines provide generic indications that can be difficult to apply to some specific subgroups of cancer patients. Knowledge of findings regarding infections in various clinical settings, according to the type of malignancy and the type of treatment, may be relevant to define targeted diagnostic and prophylactic strategies to fight infections in cancer patients. It is time to focus clinical research regarding infectious complications on specific cancer populations, in order to provide useful information for clinical practice. ■
