Risk Factors for Infection

Children with cancer are predisposed to infection for several reasons. The most common and important risk factor is neutropenia caused by myelosuppressive chemotherapy and cancer itself. Greater depth and duration of neutropenia are directly related to the risk of invasive bacterial and fungal infection and infectious mortality. Corticosteroids are commonly used as anticancer therapy and as supportive care. The negative consequences of corticosteroids on the host immune system predispose to infections such as Pneumocystis jiroveci, other fungi, and bacteria and may be an independent risk factor for sepsis and infectious mortality. Intensive chemotherapy also predisposes to infection through disruption of important protective anatomic barriers. The development of mucositis after chemotherapy administration provides a portal of entry for organisms colonizing the oral cavity and is a particularly important risk factor for viridans group streptococcal bacteremia. Likewise, compromise of the gastrointestinal mucosa may clinically manifest as enteritis, typhlitis, and colitis and can facilitate translocation of enteric pathogens to the bloodstream. Beyond chemotherapy and corticosteroids, certain procedures performed in this patient population may predispose to infection. In many countries central venous lines (CVLs) are routinely used in patients with cancer, resulting in a risk for catheter-related bloodstream and local CVL site infections. Additionally, many children require surgery for various reasons, including tumor resection, and such procedures predispose to surgical site infection. Finally, children who receive allogeneic hematopoietic stem cell transplantation (HSCT) have a unique predisposition to infections related to graft-versus-host disease (GVHD) and therapies used to prevent and treat this complication of transplantation.

Infection Classification

Infections can be classified by etiology as bacterial, fungal, viral, or protozoal, and these pathogens may infect sterile or nonsterile sites. The overall risk of infection varies depending on many factors, with intensity of chemotherapy being very important. Children who receive the most intensive chemotherapy include those with acute myeloid leukemia (AML). One study demonstrated that more than 60% of children with AML have at least one microbiologically documented infection during each phase of therapy, and the cumulative risk of infection-related mortality was 11 ± 2%. In a subsequent pediatric AML analysis, more than 80% of children experienced at least one sterile site bacterial infection and 14% experienced at least one sterile site fungal infection throughout chemotherapy. The risk of sterile site bacterial infection was 30% to 60% per chemotherapy course. In contrast, some children who receive non-myelosuppressive therapy such as those receiving maintenance therapy for acute lymphoblastic leukemia (ALL), have a very low risk of sterile site bacterial or fungal infections.

Infections may also be classified by type of infection: microbiologically documented, clinically documented, or fever of unknown origin (FUO). In an analysis of children with cancer or recipients of HSCT presenting with FN, 80% of episodes were ultimately classified as FUO, 13% had a microbiologically confirmed etiology, 6% were clinically documented infection, and 2% were invasive mycosis. The distribution of the etiology of FN will change depending on the presence of the aforementioned risk factors, especially the intensity of chemotherapy administered.

Intervention Approaches

The major approaches to reduction of infectious morbidity and mortality are prophylactic, empiric, and definitive treatment approaches. With prophylaxis, antimicrobials are administered before onset of fever or other clinical signs of infection. Empiric therapy consists of the administration of antimicrobial agents with early signs of infection, such as fever. Definitive treatment consists of administering specific antimicrobials once an infection has been documented.

Risk Factors for Infection

Children with cancer are predisposed to infection for several reasons. The most common and important risk factor is neutropenia caused by myelosuppressive chemotherapy and cancer itself. Greater depth and duration of neutropenia are directly related to the risk of invasive bacterial and fungal infection and infectious mortality. Corticosteroids are commonly used as anticancer therapy and as supportive care. The negative consequences of corticosteroids on the host immune system predispose to infections such as Pneumocystis jiroveci, other fungi, and bacteria and may be an independent risk factor for sepsis and infectious mortality. Intensive chemotherapy also predisposes to infection through disruption of important protective anatomic barriers. The development of mucositis after chemotherapy administration provides a portal of entry for organisms colonizing the oral cavity and is a particularly important risk factor for viridans group streptococcal bacteremia. Likewise, compromise of the gastrointestinal mucosa may clinically manifest as enteritis, typhlitis, and colitis and can facilitate translocation of enteric pathogens to the bloodstream. Beyond chemotherapy and corticosteroids, certain procedures performed in this patient population may predispose to infection. In many countries central venous lines (CVLs) are routinely used in patients with cancer, resulting in a risk for catheter-related bloodstream and local CVL site infections. Additionally, many children require surgery for various reasons, including tumor resection, and such procedures predispose to surgical site infection. Finally, children who receive allogeneic hematopoietic stem cell transplantation (HSCT) have a unique predisposition to infections related to graft-versus-host disease (GVHD) and therapies used to prevent and treat this complication of transplantation.

Infection Classification

Infections can be classified by etiology as bacterial, fungal, viral, or protozoal, and these pathogens may infect sterile or nonsterile sites. The overall risk of infection varies depending on many factors, with intensity of chemotherapy being very important. Children who receive the most intensive chemotherapy include those with acute myeloid leukemia (AML). One study demonstrated that more than 60% of children with AML have at least one microbiologically documented infection during each phase of therapy, and the cumulative risk of infection-related mortality was 11 ± 2%. In a subsequent pediatric AML analysis, more than 80% of children experienced at least one sterile site bacterial infection and 14% experienced at least one sterile site fungal infection throughout chemotherapy. The risk of sterile site bacterial infection was 30% to 60% per chemotherapy course. In contrast, some children who receive non-myelosuppressive therapy such as those receiving maintenance therapy for acute lymphoblastic leukemia (ALL), have a very low risk of sterile site bacterial or fungal infections.

Infections may also be classified by type of infection: microbiologically documented, clinically documented, or fever of unknown origin (FUO). In an analysis of children with cancer or recipients of HSCT presenting with FN, 80% of episodes were ultimately classified as FUO, 13% had a microbiologically confirmed etiology, 6% were clinically documented infection, and 2% were invasive mycosis. The distribution of the etiology of FN will change depending on the presence of the aforementioned risk factors, especially the intensity of chemotherapy administered.

Intervention Approaches

The major approaches to reduction of infectious morbidity and mortality are prophylactic, empiric, and definitive treatment approaches. With prophylaxis, antimicrobials are administered before onset of fever or other clinical signs of infection. Empiric therapy consists of the administration of antimicrobial agents with early signs of infection, such as fever. Definitive treatment consists of administering specific antimicrobials once an infection has been documented.

Bacteria

Bacteria are the most common cause of invasive infection in pediatric cancer patients. Over the last three decades, there has been a shift from gram-negative organisms being responsible for most infections in patients with cancer to gram-positive infections predominating. The shift toward gram-positive agents has been attributed to the wide-spread use of CVLs, chemotherapy regimens associated with mucositis, and routine use of antibiotics with gram-negative activity. The most common gram-positive infections are coagulase-negative staphylococci, viridans group streptococci, enterococci, and Staph­ylococcus aureus . The major types of gram-negative infections are Escherichia coli, Klebsiella species, and Pseudomonas aeruginosa . Bacterial sterile site pathogens are most commonly tested for and isolated from blood culture, followed by urine and, less frequently, cerebrospinal fluid.

Fungi

Pediatric patients requiring intensive chemotherapy treatment for cancers such as AML, relapsed ALL, and allogeneic HSCT have the highest risk for invasive fungal infection (IFI) because of the resultant depth and duration of neutropenia. Additional IFI risk factors include mucositis, corticosteroid use, and antibiotic exposure. The most common fungal infections in pediatric cancer are Candida spp . and Aspergillus spp. Sterile site candidal infections typically occur in the blood and urine. In contrast, invasive aspergillosis typically involves the lungs, sinuses, gastrointestinal tract, and brain. Emerging fungal pathogens that are increasingly being encountered in pediatric oncology are Mucorales (formerly referred to as zygomycetes) and Scedosporium spp. These are particularly concerning because they tend to be fatal and are challenging to diagnose antemortem.

P. jirovecii is a yeastlike fungal species that classically causes pneumonia in children with acute leukemia who do not receive P. jirovecii pneumonia (PCP) prophylaxis. In addition to chemotherapy exposure, corticosteroid use and age younger than 2 years are also risk factors for PCP. PCP prophylaxis is considered standard care for many pediatric malignancies, including acute leukemia and HSCT recipients. Trimethoprim-sulfamethoxazole (TMP-SMX) is the prophylactic drug of choice. Although the optimal dosing regimen is not clear, administration regimens of 2 or 3 days per week are most common and have been shown to be reasonably effective. Alternative regimens for allergy or other toxicities include oral dapsone, oral atovaquone, and inhaled pentamidine. Intravenous pentamidine is often used in children younger than 5 years who cannot tolerate the other options. However, clinicians should be aware that the effectiveness of intravenous pentamidine is questionable.

Viral Infections

There are many viruses that can result in infections in pediatric cancer patients. The list of potential agents continues to evolve with improved diagnostic abilities for previously unrecognized pathogens such as human metapneumovirus (HMPV). In one pediatric cancer series, respiratory syncytial virus (RSV) (31%) and rhinovirus (23%) were the most frequently detected respiratory viruses, followed by parainfluenza (12%) and influenza A (11%). RSV usually does not cause life-threatening infections in children with cancer. However, in patients with AML and HSCT recipients who do acquire RSV, infection may progress to lower respiratory tract involvement. In this setting RSV infection is associated with a 14% case fatality rate in patients with AML and a 50% case fatality rate in pediatric HSCT recipients. Adenovirus may also cause serious infection, particularly in HSCT recipients, and is a frequent cause of death in this setting. Other viral infections, including varicella-zoster virus (VZV), influenza, and cytomegalovirus, are typically not life threatening in pediatric patients with cancer who are receiving less intense therapy, but they may cause severe infection and infectious mortality in the most intensively treated children. Conversely, some infections, such as severe acute respiratory syndrome (SARS), can cause fatal infection in immunocompetent patients but is rarely associated with severe illness in pediatric cancer. It is hypothesized that an intact immune system is responsible for severity of illness and fatality in SARS, thus explaining this apparent paradox.

Protozoal Infections

In general, protozoal infections are not a common cause of fever in children with cancer in high-income countries, although these pathogens may be more prominent in some LMICs.

Initial Risk Stratification

Children with FN are not a homogenous group. Determining which children are at lower risk of complications can allow for a reduction in the intensity of therapy and monitoring. Conversely, identifying children at higher risk of complications can allow for escalation of therapy and closer observation.

In adults with FN, the risk stratification schema from the Multinational Association of Supportive Care in Cancer (MASCC) has been validated and is widely accepted as a standard approach. However, the MASCC score cannot be applied to children because age below 60 years and absence of chronic obstructive pulmonary disease are two items in the score and these are not applicable to pediatric patients. At least 25 studies of risk prediction have been conducted in pediatric cancer. These studies have been highly variable using different pediatric cancer populations and different endpoints (e.g., bacteremia, “serious infection,” death, and intensive care unit admission), making the results more difficult to interpret.

Although six of these low-risk stratification schemas have been validated in children ( Table 67-1 ), identification of a single schema to be applied across all clinical scenarios has not been feasible, likely because of the divergence in clinical settings. Therefore clinicians should review the six validated low-risk stratification schemas, choose which schema matches their clinical setting, and determine whether the application of that schema is feasible for their center. For example, some rules use biomarkers such as C-reactive protein and would require expeditious analysis and return of results to be useful. Others require clinical judgment and thus require trained health care professionals to be readily available. There are no validated risk stratification schemas to identify the pediatric high-risk patient with FN.

Initial Investigations

At initial presentation, an evaluation for the cause of fever should be conducted. The evaluation begins with a careful history and physical examination, with the assessment focusing on potential sites of infection. Sites that merit particular attention are the mouth to search for herpes simplex virus stomatitis and dental abscess, the exit site and tunnel of CVLs, perianal region for cellulitis and perirectal abscess, and fingers and toes for paronychia.

The standard evaluation should include blood cultures, with further diagnostic testing dictated by clinical signs and symptoms. Blood cultures should be obtained from each lumen of the CVL if present. The utility of a peripheral blood culture at the initial evaluation of FN is controversial. There are seven studies that evaluated the contribution of peripheral blood cultures in addition to CVL cultures in adults and children with cancer or undergoing HSCT. When these results were combined, 13% (95% confidence interval [CI] 8% to 18%) of all true bacteremias were identified only by the peripheral blood culture. This result is likely explained by false-negative CVL blood cultures when obtained culture volumes are small. However, it is not known how failure to identify these episodes of bacteremia affect patient outcomes or whether this problem can be overcome by optimizing CVL blood culture volumes. Consequently, the role of peripheral blood culture remains uncertain.

Urinalysis and urine culture testing for the detection of urinary tract infections (UTIs) is also considered controversial in the setting of FN. In immunocompetent children, a UTI is often diagnosed when there is both evidence of pyuria on urinalysis and a positive culture. The state of neutropenia limits the ability to use pyuria as a diagnostic criterion, and reliance on nitrite testing is not ideal because the test may be negative in younger children with UTI. Therefore urinalysis testing has limited utility in the setting of FN. It may be reasonable to obtain a sterile urine culture and define the presence of a UTI on the culture results alone. However, sterile urine collection can be difficult in young children and infants and often requires urinary catheterization, a procedure that is seldom recommended for a neutropenic patient. Therefore where a clean-catch or midstream urine can be collected, the results of urinalysis and urine culture should be included in the initial FN investigations. It is important to remember that antibiotic administration should not be delayed to obtain a urine sample.

There are have been four studies, including 540 episodes of FN, that investigated the impact of chest radiography (CXR) as a component of the initial FN evaluation. These studies indicate that pneumonia is detected in fewer than 5% of children without respiratory symptoms and that the omission of routine CXR does not lead to adverse outcomes. Thus routine CXR should not be performed in asymptomatic children with FN.

Initial Antibiotic Therapy

In general, empiric antimicrobial therapy choices should provide good coverage for gram-negative organisms with some gram-positive coverage. In particular, for high-risk patients with FN, antibiotic therapy should include coverage for viridans group streptococci and P. aeruginosa . Various factors must be considered when determining an empiric antibiotic regimen, route of administration, and location of therapy. These factors include patient and family social factors, clinical presentation, ability of the hospital to support an ambulatory approach, availability and cost of drugs, and local hospital resistance patterns.

The role of combination antibiotic therapy versus monotherapy for FN has been debated. Monotherapy was supported by two metaanalyses that compared monotherapy versus an aminoglycoside-containing regimen in FN and in immunocompromised patients with sepsis. These analyses demonstrated that monotherapy is not inferior and is less toxic than combination therapy. The meta-analysis in FN observed fewer treatment failures with monotherapy (odds ratio [OR] 0.88, 95% CI 0.78 to 0.99) but included only four trials that enrolled patients younger than 14 years of age. In the pediatric setting monotherapy was supported by a meta-analysis that found similar clinical outcomes when antipseudomonal penicillin monotherapy was compared to antipseudomonal penicillin plus an aminoglycoside. Therefore monotherapy is suggested for FN in pediatric patients who are clinically stable and who are treated at centers with a low rate of resistant pathogens.

Monotherapy regimens that have been evaluated in children include antipseudomonal penicillins such as piperacillin-tazobactam and ticarcillin-clavulanic acid, antipseudomonal cephalosporins such as cefepime, and carbapenems such as meropenem or imipenem. In two pediatric-specific evaluations, treatment failure, mortality rates, and adverse effects were similar when antipseudomonal penicillins were compared with antipseudomonal cephalosporins or carbapenems. There are potential downsides of carbapenems and cefepime. Carbapenems were associated with more pseudomembranous colitis compared with other beta-lactam antibiotics in a large meta-analysis. Cefepime was associated with increased all-cause mortality in another large meta-analysis when compared to other beta-lactam treated patients. However, these findings were not replicated in other studies. Consequently, cefepime remains an option for empiric therapy. Ceftazidime monotherapy lacks adequate coverage against viridans group streptococci and resistant gram-negative organisms and thus should not be used if these organisms are of concern.

Empiric glycopeptides (e.g., vancomycin) are not routinely recommended. A largely adult meta-analysis of 14 randomized controlled trials (RCTs) illustrated that inclusion of a glycopeptide did not lead to a difference in success (if addition of glycopeptide in the control arm was not considered failure) but was associated with more adverse effects. Empiric glycopeptides should be reserved for patients who are clinically unstable or who have signs or symptoms suggestive of a gram-positive infection.

Considerations for Low-Risk Patients with Fever and Neutropenia

Patients with FN who are at lower risk of adverse outcomes may be appropriate for a reduction in therapy intensity. Two strategies commonly considered are outpatient management and oral antibiotic administration. These two strategies are commonly used together, and in adults with low-risk FN, outpatient management with oral antibiotics is recommended for selected patients. Some experts have expressed concern that these recommendations may not be generalizable to children. However, over the last several years data have emerged related to efficacy and safety, costs, QoL and preference considerations for different management strategies in pediatric FN.

Modification of Empiric Antibacterial Therapy

After therapy for FN has been initiated, the initial empiric regimen should be adjusted to provide appropriate coverage for any positive microbiology results or identified clinical focus of infection. In patients in whom empiric glycopeptides or dual gram-negative coverage was initiated, reassessment at 24 to 72 hours should be conducted. In the absence of a specific microbiological reason to continue these agents, they should be discontinued. In the event of persistent fever, careful evaluation for an undetected source of infection is important. In this setting modification of antibiotics, including addition of empiric vancomycin, is not warranted in children who remain clinically stable. Children who clinically deteriorate warrant broadening of empiric antibacterial therapy to include coverage for resistant gram-positive, gram-negative, and anaerobic organisms.

Empiric antibiotics should be discontinued if cultures are negative, the child is clinically well, fever has resolved, and there is evidence of neutrophil recovery. One randomized trial of pediatric low-risk patients found that cessation of antibiotics on day 3 irrespective of count recovery versus continuation of antibiotics was associated with similar outcomes. However, Enterobacter bacteremia occurred in one child in the early cessation study arm. Consequently, it is reasonable to discontinue antibiotics on day 3 in low-risk children with FN who have become afebrile with negative cultures as long as careful monitoring is in place. In high-risk patients the optimal duration of antibiotic therapy is unknown in the setting of persistent profound neutropenia. A small study of 33 high-risk patients suggested that cessation of empiric antibiotics on day 7 is associated with bacteremia and poor infection outcomes compared with continuation for 14 days. Thus continuation of empiric antibiotics for at least 14 days for high-risk FN in the absence of evidence of neutrophil recovery is a reasonable strategy.

Evaluation for Invasive Fungal Infection and Empiric Antifungal Therapy

Children with FN and persistent or recurrent fever that persists 96 hours or more after initiation of broad-spectrum antibiotics and who are at higher risk of IFI should undergo an evaluation for fungal infection, including careful physical examination, blood and urine cultures, and computerized tomography (CT) of the chest. The role of routine CT sinuses and imaging of the abdomen have not been defined in the standard investigation of IFI, although routine CT sinuses should not be conducted in children 2 years of age or younger because of insufficient pneumatization of the sinus cavities. In children with demonstrated pulmonary lesions, investigations may include bronchoalveolar lavage and lung biopsy, although fatal bleeding may occur with biopsy of any angioinvasive mold lesion. Thus the decision to biopsy requires careful consideration.

Empiric antifungal therapy should consist of either caspofungin or liposomal amphotericin B (L-AmB) because these two therapies are similarly effective and L-AmB is slightly better and less nephrotoxic than amphotericin B deoxycholate. Empiric antifungal therapy may be discontinued at resolution of severe neutropenia if the patient is clinically well without evidence of an IFI.