Rudolph's Pediatrics, 22nd Ed.

CHAPTER 229. Fever and Infection in the Immunocompromised Patient

David B. Haslam and Jeffrey Scott McKinney

A key element in the effective care of immuno-compromised patients with fever and infection is hypervigilance. Not only are these patients at increased risk from a diverse range of microorganisms, but also their resulting infections can present with subtle or atypical symptoms and can progress to difficult-to-treat chronic disease states, or to acute life-threatening clinical decompensations. Given these facts, empiric anti-infective therapy is sometimes justified before the definitive cause of a fever is elucidated. Yet, the same facts also argue for the vital importance of making an accurate and exact diagnosis, to guide definitive ongoing therapy.

Toward this end, the general features of a thoughtful diagnostic approach to fevers in immunocompetent patients remains valid for immunocompromised patients, including a systematic consideration of noninfectious diagnoses (see Chapter 228). Likewise, a thorough exposure history remains essential and is expanded to include microorganisms traditionally considered to be “environmental” or of “low virulence.”

Published guidelines for the use of antimicrobial agents in certain sets of immunocompromised patients outline contemporary standards of care and can provide useful algorithms to help improve outcomes for these patients (for clinical practice guidelines, refer to However, a key first principle of such guidelines should be the importance of frequent reevaluation of an immunocompromised patient’s course and care, given his or her extreme medical vulnerability. Furthermore, optimal care may require input from a physician with expertise in treating immuno-compromised patients infected with microorganisms that are often difficult to treat, such as fungi or mycobacteria.


Defined defects in particular immunologic functions (see Chapters 187 and 188) correspond with increased patient risk for specific infections (Table 229-1). These immune defect–specific risks offer basic principles and insights to influence diagnostic workup and empiric therapy for children known to have a particular type of immunocompromise. Conversely, certain infectious disease presentations provide clues to trigger and guide an evaluation to first diagnose specific immune defects in children.


Aside from inherited immunodeficiencies, acquired conditions such as HIV/AIDS and severe malnutrition are also significant root causes of immunodeficiency world-wide, particularly among children in medically under-served and economically depressed areas.

In developed countries, immunodeficiency is now most commonly iatrogenic in origin. Cytotoxic chemotherapy for malignancy is an extensively studied, and repeatedly encountered, iatrogenic cause of immune dysfunction. Moreover, severe connective tissue disease and increasingly common organ transplantation now provide other medical indications for immune suppression, often mediated by newer drugs or regimens with novel immunosuppressive effects. As a consequence, both the incidence of severe immunodeficiency and the variability of its manifestations have markedly increased over previous decades.

As shown in Table 229-1 and discussed in Chapter 187, T-cell–mediated immunity is particularly important for defense against fungal and viral infections, the reticuloendothelial function of the spleen for clearing encapsulated bacteria from the blood, terminal complement components for lysing opsonized bacteria such as Neiserria meningitidis, the neutrophil oxidative burst for killing catalase-positive bacteria and fungi within neutrophils, and the importance of antitoxin antibodies in preventing toxic shock induced by the TSST-1 superantigen.

Table 229-1. Immunologic Functional Defects and Susceptibility to Infections

Yet, from a clinical perspective, the pathophysiology of infections in immunocompromised hosts often reflects a combination of factors. For example, following cytotoxic chemotherapy for cancer, patients are predisposed to overwhelming sepsis not only because they are neutropenic, but also as a function of their mucosal breakdown, intravascular catheter infection risks, and prior antimicrobial therapy that can select for drug-resistant microbial flora.

Patients with acquired immunodeficiency are clinically distinct from those with primary immunodeficiency for several reasons. As described previously, acquired immunodeficiency states often involve multiple, compounded, risks for infection. However, the root causes of acquired immunodeficiency, whether secondary to immunosuppressive drugs or HIV infection, can sometimes be clinically modified by alterations in therapy. Accordingly, intentionally reducing immunosuppressive drug dosing (or, as relevant, HIV viral load) can be a key element of definitive anti-infective therapy.


Management of fever in the child with profound neutropenia provides one paradigm for the care of suspected infection in a child with severe immunocompromise. Obviously, this approach is not universally applicable to all immunocompromised children, but it includes core concepts that can be generally applied to children with immune deficit. These concepts include the association between certain defects of immunity and propensity for specific infections, empiric therapy choice and modification, step-wise diagnostic evaluation, treatment duration decisions, and interval assessment parameters.


In the context of antineoplastic therapy, neutropenia is traditionally defined as an absolute neutrophil count (ANC) <500 cells/mm3, or as an ANC of 1000 cells/mm3 with a predicted decrease to <500 cells/mm3. Quite distinct from the prolonged fever duration required for classic definitions of fever of unknown origin (FUO), fever is here defined as a single oral temperature ≥38.3°C (101°F), or ≥38.0°C (100.4°F), for at least 1 hour.1

Key clinical features of febrile neutropenic patients depend on the depth and duration of their neutropenia. Hence, a detailed history of recent cytotoxic treatments, to predict an anticipated ANC nadir, can be particularly helpful. Fifty percent of febrile neutropenic patients prove to have infections, with risk of infection increasing with decreased ANC and with duration of low ANC.1,2 One in five patients with fever and an ANC <100 have bacteremia; of particular relevance are the bacteria listed in Table 229-1, but an extremely diverse range of other bacteria can also be responsible.1 Neutropenia also predisposes to severe fungal infections, especially with Candida and Aspergillusspecies, particularly among patients who have received courses of broad-spectrum antibacterial drugs.


Neutopenic hosts often lack classic signs and symptoms of inflammation, such that severe infections can exhibit remarkably occult presentations. Furthermore, lack of classic inflammation on exam does not preclude rapid progression to systemic decompensation. Indeed, a fundamental challenge in caring for immunocompromised patients is that they can present in extremis from overwhelming infection. Accordingly, attention to signs of shock, hypoxia, or deep-organ infections is essential.

Examination should seek even subtle signs of infection, such as pain in anatomic sites, including the periodontium, pharynx, esophagus, lung, perineum, eye, skin and nails, and near vascular access sites. Imaging studies should include a chest X-ray for patients with any respiratory signs or symptoms, or for those who will be managed as outpatients. CT or MRI of the abdomen, sinuses, or other sites may also be recommended, based on relevant symptoms or previous history of infection at these sites.

Blood cultures, including those drawn from all lumens of central lines, are a cornerstone of diagnostic evaluation to detect bacteremia or fungemia, identify specific microorganisms, and define their susceptibilities to antimicrobials. Quantitative cultures, drawn in parallel of blood from central line lumens versus peripheral blood, can help assess the likelihood of a central line serving as an ongoing nidus of infection.  Rapid microbiologic stains of any exudates, biopsy specimens, or positive cultures can provide important early insights for directed therapy. Importantly, evidence of deep infections on imaging often necessitates direct biopsy and culture for definitive diagnosis and therapy.

Additional diagnostic tests traditionally include urinalysis, complete blood count with differential, blood tests for liver and renal function, electrolytes, and, in some cases, cerebrospinal fluid evaluation for potential intracranial infection.


Regardless of findings on physical examination and initial laboratory investigations, all neutropenic patients with fever should receive empiric antimicrobial therapy directed at the microorganisms most commonly identified in this setting. Whereas gram-negative rods (particularly Escherichia coli, Klebsiella, and Pseudomonas) were formerly the most commonly isolated, more recent trends have shown gram-positive organisms accounting for more than 50% of blood cultures in children undergoing chemo-therapy.6,7 Coagulase-negative staphylococci, which are the single most commonly isolated organisms, follow a relatively indolent course. However, some gram-positive bacteria, such as Staphylococcus aureus and the viridans Streptococcus mitis, may be associated with rapidly progressing shock, noncardiogenic pulmonary edema, and multiorgan failure, which is more traditionally associated with gram-negative sepsis.8

Empiric parenteral antimicrobial therapy can take one of several forms, but needs to address contemporary local patterns of infection and resistance. As a general rule, a regimen should provide broad coverage for gram-negative organisms, including Pseudomonas, and at least some coverage for S aureus. US Food and Drug Administration–approved single-agent regimens meeting these criteria include ceftazidime, cefepime, or carbapenem (eg, meropenem or imipenem). A carbapenem is preferred if local rates of extended-spectrum β-lactamase (ESBL) positive strains are high, or if Klebsiella infection is suspected.

Although initial therapy with two agents is often used, studies to date do not detect marked differences in outcome between empiric mono-therapy and duotherapy. The potential benefits of adding an aminoglycoside to the single-agent regimens listed previously include synergistic activity against many gram-negative organisms and a decrease in the emergence of drug resistance. However, this must be balanced with known risks for aminoglycoside-induced ototoxicity and nephrotoxicity. These risks are not only acute and dependent on renal function dynamics,9 but they may also be cumulative among repeatedly exposed patients.10

The decision whether vancomycin should be included in the initial regimen deserves particular attention. Unwarranted overuse of vancomycin has significant negative effects, including powerful selection for vancomycin-resistant enterococcus. Past recommendations from the Infectious Diseases Society of America (IDSA) have concluded that vancomycin should only be added to the initial regimen if the patient has an obvious catheter-related infection, severe mucositis, β-lactam–resistant Streptococcus pneumoniae, hypotension or other signs of sepsis, or known colonization with methicillin-resistant Staphylococcus aureus (MRSA). The full clinical relevance of what appears to be a new epidemiologically ascendant clone of MRSA remains to be determined, but it may well be a serious additional gram-positive threat to immunocompromised children .

FIGURE 229-1. Guide to initial management of febrile neutropenic patient. See text for discussion regarding choice of antibiotics in monotherapy and duotherapy. MRSA, methicillin-resistant Staphylococcus aureus.

The empiric duotherapy of vancomycin and ceftazidime regimen has been extensively studied in neutropenic patients and found to be safe and efficacious. Vancomycin should be discontinued after 48 hours if the clinical course is stable and gram-positive organisms are not isolated.

At approximately 3 to 5 days of therapy, the condition of the patient is the primary determinant of further therapy (Figs. 229-1 to 229-3) and may allow for a change to an appropriate oral antibiotic such as cefixime. Such a course assumes absence of signs of sepsis on admission, no evident source of infection, clinical improvement, negative cultures after 48 hours, and an absolute neutrophil count (ANC) > 100/mm3. Alternately, such patients may be discharged from the hospital to continue parenteral antibiotics at home. In patients who remain afebrile after the third to fifth day, antibiotics may be discontinued after 7 days, particularly if the ANC is >500/mm3 or the bone marrow is recovering. Patients in whom the ANC remains <100/mm3, or those with ongoing mucositis or other factors predisposing to recurrent infection should continue parenteral antibiotics as an outpatient, even if they become afebrile.

Some centers have adopted guidelines that allow outpatient management of selected low-risk children with fever and neutropenia after a single dose of parenteral antibiotics and careful evaluation.11,12 After obtaining a complete blood count and blood cultures, the child is given a parenteral antibiotic such as ceftazidime, then discharged on an oral antibiotic, such as ciprofloxacin, with close follow-up and careful reevaluation. This initial outpatient management is usually not considered for patients with hemodynamic instability, severe mucositis, or those undergoing induction chemotherapy or bone marrow transplantation.

The median time to fever defervescence in adequately treated patients is 5 days.13,14 Accordingly, if a patient remains febrile but is otherwise clinically stable, and if a definitive infecting organism is not identified on early cultures, the clinician may wait 5 days to make changes in the antimicrobial regimen. Among neutropenic patients whose fevers continue more than 5 days, the following possibilities should be actively considered: (1) a nonbacterial infection, (2) a drug-resistant organism, (3) an infection poorly accessible to antibiotics (eg, an abscess or catheter-associated infection), (4) inadequate dosing of antimicrobials, or (5) a noninfectious source of fever. To assess these possibilities, patients should be examined thoroughly for presence of a previously unidentified focus of infection. Blood cultures, including fungal cultures, should be repeated. Relevant drug serum levels should be obtained. Imaging studies of the abdomen, chest, or sinus should be considered. Depending on the condition of the patient, three treatment options are available for fevers that do not respond to initial empiric therapy as shown in Fig. 229-3.

FIGURE 229-2. Management of patients who become afebrile in first 3 days of initial antibiotic therapy. *, Clinically well; **, absolute neutrophil count <100/mm3, severe mucositis, or unstable vital signs.

The total duration of antimicrobial therapy in febrile neutropenic patients depends on neutrophil count, resolution of fever, and presence or absence of a focus of infection. As described previously, patients who become afebrile and have an ANC > 500/mm3 (or who have a compelling increase in their ANC and no mucositis) may have antibiotics discontinued after 7 days. In contrast, persistent fever in patients with ongoing neutropenia requires antimicrobial therapy (including antifungal therapy) for a minimum of 2 weeks. After this course, imaging studies of the chest, abdomen, and other sites should be considered, if not already performed. If imaging studies are normal and the patient is febrile but clinically stable, consideration may be given to discontinuing antimicrobials after 2 weeks (or, alternatively, 4–5 days after the ANC reaches 500/mm3), with ongoing close patient evaluation.

FIGURE 229-3. Management of patients who have persistent fever after 3 days of treatment and in whom etiology of fever is unknown. See text for discussion of adjustment in antibiotics or the addition of amphotericin B.

As described previously, neutropenia is a key predictor of risk for infection, and absolute neutrophil count criteria help guide therapeutic decisions. Yet, paradoxically, neither granulocyte infusions nor granulocyte colony-stimulating factors (G-CSFs) are recommended as routine treatment for fever and neutropenia.15 Although G-CSF can shorten the duration of neutropenia, it does not decrease mortality or shorten duration of fever. Accordingly, G-CSF is used primarily in severely neutropenic patients with documented infections that do not respond to appropriate therapy, or when a prolonged delay in marrow recovery is anticipated.


Febrile patients who are solid organ transplant recipients offer contrasts to those with fever and neutropenia, and illustrate some additional general principles. Among transplant patients, lymphocyte-based immune function is most compromised, and susceptibility to different infections varies as a function of time.16


Early following transplantation (eg, <1 month), nosocomial infections predominate, including aspiration pneumonias, and catheter, wound, or anastomotic leak infections. Recipient-derived infections are most common, which reflect host microbial colonization (eg, with Aspergillus, Pseudomonas, or resistant species such as methicillin-resistant S aureus, vancomycin-resistant enterococcus, or nonalbicans Candidaspecies). Less commonly, acute donor-derived infections (eg, herpes simplex virus [HSV], West Nile virus, Trypanosoma cruzi, HIV, lymphocytic choriomeningitis virus, rabies) can also occur.

In the intermediate posttransplantation period (eg, 1–6 months), immunosuppression is often at its most intense. During this time, there are notably high risks for pneumocystis pneumonia and herpes simplex virus (HSV), varicella zoster virus, cytomegalovirus (CMV), and Epstein-Barr virus (EBV), risks that can be attenuated by prophylaxis with trimethoprim-sulfamethoxazole and oral antivirals, respectively. Other opportunistic infections classically seen during this stage include listeria, nocardia, toxoplasma, Mycobacterium tuberculosis, hepatitis B and C viruses, adenovirus, influenza, BK polyomavirus, Cryptococcus neoformans, and endemic fungi such as histoplasmosis. In relevant geographic contexts, strongyloides, Leishmania, and Trypanosoma cruzi infections can also occur.

In the later posttransplant period (eg, after 6 months), immunosuppressive medication is traditionally decreased, but host susceptibility remains significantly increased for listeria, nocardia, invasive fungal disease, and some otherwise unusual organisms such as Rhodococcus species. Likewise, late viral infections can occur with CMV, HSV, and JC polyomavirus. Transplant-associated malignancies such as posttransplantation lymphoproliferative disorder (often EBV associated and reported in more than 2% of pediatric lung transplant recipients)17 and late allograft rejection can also present with fever during this time.

As for patients with neutropenia, the degree of immunosuppression in transplant patients is believed to influence risk for infection. However, clinical indices of this immunosupression are less direct than the quantitative measurements such as the absolute neutrophil count for patients with neutropenia, or the CD4 count for patients with HIV. Thus, serum levels of immunosuppressive drugs and judgments about the relative likelihoods of allograft rejection or host infection are used to titrate immunosuppressive therapies. Some viral infections seen in transplant patients (eg, CMV, EBV, adenovirus, BK) can now be assessed using quantitative polymerase chain reaction (PCR)–based tests for viral load; these too, can help guide decisions about anti-infective and immunosuppressive therapy.


Complications of infections in immunocom-promised hosts often relate to three core issues: (1) initial diagnoses can be missed, secondary to radically altered features of disease presentation as compared to those observed in immunocompetent hosts; (2) even traditionally “nonpathogenic” organisms can cause severe disease in immunocompromised hosts; and (3) antimicrobial agents may not adequately arrest infectious disease progression. These concerns reiterate the general importance of maintaining a broad differential diagnosis for possible infectious diseases, in which almost any microorganism isolated from what should be a sterile body site should be respected as a potential pathogen. Accordingly, direct biopsy and cultures of potential foci of infection are extremely important.

Prognosis can be improved with enhanced patient and physician awareness of, and rapid response to, specific infection risks, as exemplified by improvements in early recognition of and intervention for possible bacteremia in patients with sickle cell disease.19


To decrease the likelihood of infection in immunocompromised children, preventive strategies include prophylactic antimicrobials, vaccination, passive immunization via immunoglobulins, and attempts to minimize exposures to known infectious disease risks.

The use of prophylactic antimicrobials are indicated, in some well-studied contexts. Trimetho-prim-sulfamethoxazole prophylaxis significantly decreases risk of pneumocystis pneumonia, and has proven benefit in patients with compromised T-cell immunity, such as transplant recipients, patients with advanced HIV infection. It is also sometimes used as prophylaxis for patients with chronic granulomatous disease, primarily for its efficacy in preventing nocardia infections.20 Antifungal prophylaxis in severely immuno-compromised patients has seen recent changes, secondary to expanded antifungal drug options. Posaconazole prophylaxis, in particular, has been shown to decrease invasive fungal infections and improve mortality in patients with neutropenia,21 and is US Food and Drug Administration approved for prevention of fungal infections, including Candida and Aspergillus in patients with hematologic malignancies, bone marrow transplants, and graft-versus-host disease. Antiviral drugs against herpesviruses are often used in transplant patients, either in universal prophylaxis or so-called preemptive therapy regimens, to decrease risk of cytomegalovirus.22


Vaccination recommendations for immuno-compromised hosts are regularly updated by the Infectious Diseases Subcommittee of the American Academy of Pediatrics (for the most recent American Academy of Pediatrics Red Book, refer to Key recommendations in past guidelines have emphasized the particular benefit of supplemental vaccination against pneumococcus (now, both the conjugated heptavalent vaccine and the polysaccharide vaccine covering 23 serotypes) for patients with splenic dysfunction (eg, sickle cell disease, asplenia, or polysplenia). Patients with splenic dysfunction or complement deficiency should also receive the meningococcal vaccine, with administration relatively early in childhood. Annual influenza vaccination is recommended for immunocompromised patients in general.

Of note, vaccines composed of live viruses (eg, measles-mumps-rubella [MMR], varicella, and oral poliovirus vaccine) are often withheld from children with immunodeficiency. However, live vaccines are considered of net benefit in certain circumstances. Single-antigen varicella vaccine may be administered to HIV-infected children with >15% CD4 T cells and to susceptible children with cancer when in remission at least 3 months since their last chemotherapy.23 MMR may be administered at 12 months of age to HIV-infected children without advanced HIV disease, or sooner during a measles epidemic. The combined measles-mumpsrubella-varicella (MMRV) vaccine is not recommended for any child with known or suspected immunodeficiency.23

Passive immunization with intravenous immunoglobulin preparations may be indicated in two main contexts: (1) among patients for whom profoundly depressed levels of immunoglobulin are a root cause of their immunocompromised state, in which pooled immunoglobulin is given, essentially as a replacement factor; and (2) when pathogen-specific immunoglobulins may prevent or abrogate certain disease progression, as for varicella hyperimmune globulin administration to highly immunocompromised, varicella-susceptible hosts soon after putative virus exposure, or cytomegalovirus (CMV) hyperimmune globulin for some transplant patients with CMV infections.22

Attempts to minimize exposures to known infectious disease risks remain fundamentally important. Fastidious central line care is imperative. Likewise, immunocompromised patients should be educated on prudent avoidance of potential environmental point sources of pathogens, including unpasteurized dairy products; well and lake water, and even previously treated water in standing collections; soil (eg, building sites, caves, gardening or agricultural work); contact with feces or secretions; exposure to animals of notable zoonotic risk (eg, birds, reptiles, rodents, cats); and undercooked meats or unwashed produce. In extreme cases, however, such as for patients in bone marrow transplant units, even small breaches of draconian isolation measures have been associated with catastrophic infections. Finally, the use of broad-spectrum antimicrobials without a clear indication is also an important avoidable environmental risk. Inappropriately broad antimicrobial use is a known risk factor for subsequent serious infection with highly resistant bacteria and invasive fungi.