Consideration of treatment of the febrile neutropenic patient includes a brief summary of risk factors, organisms responsible for the disease, workup and management of a patient with respect to antimicrobial therapy, and immunoenhancement. Many clinical disease entities can cause a spectrum of immune suppression, and solid and hematologic tumors vary with respect to the degree of immune suppression they produce.
Risk factors for infection
Neutropenia is defined as an absolute neutrophil count that is less than 1000 cells/mm3. As the count falls below 1000 cells/mm3, the risk of infection increases. Although the degree and rate of fall of neutropenia are important determinants of infection, the duration of profound neutropenia (absolute neutrophil count of <100 cells/mm3) directly predicts the development of infection. Development of infection approaches 100% if profound neutropenia persists for 3 weeks or longer. Initial induction of antitumor therapy produces a period of prolonged neutropenia and subjects the patient to the greatest risk of infection. If an infection develops, its successful outcome is determined in large part by recovery of the neutrophil count, as well as the choice of antimicrobials. In addition to purely quantitative decreases, neutrophils may have qualitative defects that impair their function. The latter may be the result of premature release of the cells from the marrow secondary to invasion of the marrow by tumor, or intrinsic killing defects caused by cytotoxic agents, irradiation, or antimicrobial therapy.
A number of cell lines can be markedly affected by the same insults. Profound lymphopenia results in decreased cell- and antibody-mediated immunity. Other important risk factors for development of infection include the following: disruption of the skin and mucosal barriers secondary to the effects of chemotherapeutic agents or malnutrition; splenectomy; alteration of endogenous microbial flora secondary to antimicrobial prophylaxis or colonization with nosocomial-acquired organisms, the presence of an indwelling catheter; and exposure to health care workers, air, water, food, and supportive equipment found in the hospital (e.g., ventilators, etc.).
Organisms causing infection
The types of organisms causing infection in neutropenic patients with leukemia have changed over time, largely because of antimicrobial prophylaxis and therapy. Before such therapy, the predominant organisms that caused infections were aerobic gram-positive bacteria (e.g., frequently S. aureus). With the introduction of effective antistaphylococcal penicillins and cephalosporins and their regular use as empiric therapy in neutropenic patients, the incidence of gram-positive bacterial infections was reduced. However, an increase in gram-negative bacterial infections was subsequently noted, caused by E. coli, Pseudomonas aeruginosa, and Klebsiella species. Because of the high mortality associated with infection caused by these organisms, early intervention with broad-spectrum antibacterial agents at the first sign of fever became routine. This practice has greatly improved survival and has led to a decrease in mortality rates for infections caused by these organisms.
Most recently, gram-positive bacterial organisms including S. aureus and S. epidermidis, viridans streptococci, Corynebacterium jeikeium and Enterococcus species have again become major pathogens. This change occurred primarily because of the placement of indwelling intravenous catheters. In addition, antimicrobial prophylaxis and the reduction in use of specific antistaphylococcal agents as first-line therapy may also have contributed to the change. Fungal organisms such as Candida species and Aspergillus species have emerged as significant pathogens as well.
Organisms that cause infection in these patients can come from a newly acquired nosocomial pathogen, the patient's endogenous flora, or reactivation of latent organisms such as herpes viruses. The majority of organisms found to cause infection in neutropenic patients were from the patient's flora that had been modified by antibiotic treatment.
Although some generalizations can be made with respect to which organisms cause infections in neutropenic patients, the clinician should learn about the specific flora found in a given hospital, the organisms that are most frequently isolated from febrile neutropenic patients in that hospital, and the antimicrobial sensitivity and resistance patterns of these organisms.
Principles
Empirical treatment of immunocompromised patients requires that the practitioner know the local epidemiology of infectious diseases, their complications, and the susceptibility patterns for nosocomial pathogens.
Investigation and empirical treatment
A temperature of more than 100 °F (38 °) is usually defined as a medically important fever, and when observed in neutropenic patients should signal a search for infection.
Because of the neutropenia, an inflammatory response may not be evident, and fever may be the only clue that infection is present. Infected neutropenic patients with lung infections may have normal chest x-rays, and they may not be able to produce purulent sputum. Infected skin may not show signs of cellulitis or abscess. A careful physical examination can help to define a source of infection. Particular attention should be paid to the ocular fundi and mucocutaneous sites such as the nasooropharynx and perirectal areas — common sites that can reveal evidence of infection. In most patients no obvious source of infection is evident. Therefore appropriate cultures of blood, urine, skin lesions, throat, sputum, or cerebrospinal fluid must be obtained. Radiologic examination of the chest, sinuses, and teeth may be helpful in defining a site of infection. Unless the physical examination or x-ray films reveal a specific site of infection to guide specific antimicrobial therapy, the initiation of empirical broad-spectrum antimicrobial therapy must be instituted.
Many studies have examined the effects of single-agent versus combination therapy as empirical regimens to treat the febrile neutropenic patient. These studies have provided information on when to initiate therapy for gram-positive bacterial organisms and fungal agents, how to modify therapy and its duration, and how to treat the persistently febrile patient. Broad-spectrum combination therapy that includes an antipseudomonal β-lactam antibiotic (e.g., mezlocillin, piperacillin, or ceftazidime) and an aminoglycoside (e.g., gentamicin, or tobramycin) is recommended by many investigators. Because gram-negative bacterial organisms such as Pseudomonas species can be associated with rapid mortality, a potentially synergistic combination regimen is attractive. Such combinations provide broader empirical coverage than does therapy with only one agent. Gram-negative coverage and moderate gram-positive coverage are usually required.
The choice of which agents to use depends on the most common pathogens and their patterns of antimicrobial resistance in a given hospital. Aminoglycosides have the potential disadvantage of causing nephrotoxicity and ototoxicity, and concentrations in plasma must be monitored. Aminoglycosides should not be used as monotherapy because of insufficient efficacy. Initial empirical monotherapy with agents such as imipenem, ceftazidime, or ticarcillin-clavulanic acid produces adequate efficacy without the toxicities associated with aminoglycosides. Interpreting the results of these studies is not simple because “adequate clinical response” may not be defined and certainly is different between studies (e.g., some focus on overall clinical outcome or death versus survival, whereas others note the development of an infection or the necessity to change or add an antimicrobial agent as failure). In many of these studies, no organism was identified, with response defined as resolution of fever.
A major issue in empirical regimens involves the addition of an agent to specifically provide coverage for gram-positive cocci, especially staphylococci. Although many infections in neutropenic patients are caused by gram-positive organisms, the morbidity and mortality of infection with these organisms is markedly less than those of infections caused by gram-negative bacilli. A retrospective review concluded that the addition of vancomycin as an initial empirical agent is not necessary. Although more gram-positive infections were noted in the group of patients who did not receive vancomycin initially, successful treatment of documented infections caused by gram-positive organisms was not different whether vancomycin was instituted initially or delayed. However, if the febrile episode appears to be the result of an infectious focus with gram-positive organisms (e.g., an infected indwelling intravenous catheter), then the early empirical use of vancomycin would seem prudent and warranted.
Most investigators agree that since the majority of infections found initially in the febrile neutropenic patient are bacterial, the early use of antifungal agents is not warranted. Most patients respond to empirical antibacterial therapy within a week. By this time several issues may have become clear. The infecting organism may be identified; the neutrophil count may begin to rise; the organism may be identified but the infection may fail to respond; a site of infection may be identified but no organism is identified; or no site of infection may be identified or organism isolated and the patient may still be febrile, possibly indicating fungal infection. If no source or organism is identified, it appears prudent to withhold antifungal treatment until 5 to 7 days of empirical antibacterial therapy have been completed, because some patients may have bacterial infections that are slow to respond to treatment.
Fungal infections are extremely difficult to diagnose, but initiation of empirical therapy may prevent dissemination. Once the decision is made to initiate empirical antifungal therapy, amphotericin B is usually chosen.
The duration of therapy is a matter of some controversy. Some investigators believe that empirical antifungal therapy should be continued until neutropenia resolves. Others continue antifungal therapy based on the clinical response. If a fungal source is identified, a standard course of antifungal treatment would be continued. The experience with the imidazoles or newer triazoles in empirical antifungal therapy is limited. If the patient continues to be febrile and neutropenic despite an antibacterial and antifungal regimen, then the antifungal agent might be discontinued pending re-evaluation.
The duration of empirical antibacterial therapy is controversial. If the patient becomes afebrile while receiving an empirical regimen but remains neutropenic, the duration of treatment depends upon the clinical status of the patient and the neutrophil count. In general antibiotic therapy should be continued until neutrophil count is >500/mm3.
The above discussion has revolved around empirical therapy for the neutropenic patient with fever but with no identified site of infection or identified pathogen. After empirical therapy has been initiated, modification of a regimen may be necessary based on subsequent clinical developments. For example, if perianal tenderness, abdominal pain, necrotizing gingivitis, or mucositis occur, anaerobic organisms may be involved and coverage for anaerobic bacteria should be added. In general, antimicrobial therapy should not be directed toward anaerobes unless they are cultured from the blood or there is clinical suspicion of anaerobic infection. Radiologic evidence of diffuse interstitial pneumonitis may indicate infection with Pneumocystis carinii. If so, a trial of trimethoprim-sulfamethoxazole could be initiated. New pulmonary infiltrates or lesions developing while the patient is receiving antibacterial therapy could indicate pulmonary fungal infection. In such cases, a biopsy (either percutaneous or open) of the pulmonary lesion should be attempted. If the patient is unable to tolerate these procedures, empirical antifungal therapy should be initiated. Finally, if an organism was found initially but breakthrough bacteremia has occurred, the regimen should be modified to include an antibacterial agent not having cross-resistance.