Special Feature

Invasive Pulmonary Aspergillus in the ICU

By Stephen W. Crawford, MD

Many intensivists have infrequently diagnosed and treated patients with invasive pulmonary aspergillosis in the past. This has been a disease of immunosuppressed hosts and organ transplant units. However, there is an increasing incidence of neutropenic and immunosuppressed hosts in the intensive care unit (ICU). This is due to the rise in the use of solid organ and hematopoietic stem cell transplantation, as well as the use of more intensive chemotherapies to treat malignancies. In addition, numerous infections, drugs, and systemic diseases are associated with neutropenia, a major risk factor for invasive pulmonary aspergillosis. (See Table 1.)

    Table 1
    Some Causes of Neutropenia                       
    Drug myelosuppression
    Viral infection
    Late stages of AIDS
    Herpes group virus infection
    Congenital deficiency
    Inherited cyclic neutropenia
    Functional defects
    Chediak-Higashi syndrome
    Myeloperoxidase deficiency
    Chronic granulomatous diseases

Issues for the ICU

Invasive pulmonary aspergillosis increasingly is becoming a critical care issue since the ICU now cares for increased numbers of immunosuppressed and neutropenic patients. Invasive pulmonary aspergillosis is often a complication in immunosuppressed patients who have another critical illness. Often, invasive pulmonary aspergillosis is not the primary reason for ICU admission, and often may not be the cause of death. More troublesome to the intensivist, invasive pulmonary aspergillosis is often an unsuspected diagnosis in patients not thought to be at high risk (such as those with severe COPD or AIDS).

Numerous infections are emerging as problems, especially in patients with neutropenia. These infections include an increasing incidence of infection with Gram-positive cocci, such as Streptococcus viridans, vancomycin-resistant enterococci, and coagulase-negative staphylococci. Potentially more disturbing is an increasing problem with invasive fungi. Reports of Candida infections have increased 20-fold since the 1980s. Aspergillus infections show a 14-fold increase. In addition, there is an increasing recognition of "unusual" fungal species, such as Trichosporon, Fusarium, and Mucor.

When to Suspect Aspergillus in the ICU

Incidence and severity of fungal infection depends on the depth of neutropenia (e.g., < 500 cells/mL), as well as its duration. Increasingly, it is apparent that profound phagocyte dysfunction contributes as well. The role of phagocyte function is illustrated by the reports of nosocomial Aspergillus fungal infections in the ICU. Several centers have reported hospital-acquired invasive pulmonary aspergillosis among patients exposed to high levels of fungal spores from ventilation system sources while receiving high-dose corticosteroids, even when these drugs have been administered only for a short period. Clearly, steroid-induced neutrophil dysfunction can predispose one to invasive pulmonary aspergillosis when the exposure is sufficient.1,2

We can draw much about the risks for invasive pulmonary aspergillosis in the ICU from the experiences in neutropenic patient populations. The complications of severe, short-duration neutropenia in patients with peripheral stem cell transplantation demonstrate the protective effect of a short duration of neutropenia on invasive fungal infections. Neutropenia was nearly universal among these patients in a recent report, but lasted for less than five days. Neutropenic fever was noted in 94% and the patients defervesced in a median time of four days. Bacteremia was noted in 39% with Gram-positive cocci the predominant organism. Unlike previous reports from bone marrow transplant recipients with prolonged neutropenia, pulmonary infiltrates were detected in only 5% and no fungal infections were noted. There were no infection-related deaths.3

Pneumonia in a patient with neutropenic fever is an ominous finding. Infection is documented as the cause of neutropenic fever in less than one-third of cases. Lung and skin are the most commonly identified sites when a source of infection is localized. Involvement of these sites tends to occur days later in the fever course and portends a worse prognosis. Notably, more than half of documented lower respiratory infections are due to fungi.4 Not surprisingly, the initial antibiotic choices do not significantly affect the resolution of pneumonia in patients with neutropenic fever after chemotherapy. The reported response rates are 61-73%. Most empiric broad-spectrum antibiotic regimens are equally effective. Only 27% of these patients with pneumonia respond without the addition of antifungal agents.5 These facts strongly suggest that more than half of the cases of neutropenic fever with pneumonia have fungal infection.

Barton and Schuster recently examined the cause of fever after resolution of neutropenia due to chemotherapy. They observed that 20% of patients developed fever after the blood leukocyte count had returned to normal. The ascribable etiologies were "unknown" or "noninfectious" in 62%. Bacteria accounted for 17% and fungal for 21%. Among the fungal infections, one-third were Candida species and two-thirds were Aspergillus species.6

Table 2 summarizes the key points in considering infection in neutropenic patients.

    Table 2
    Key Points About Infection in Neutropenia                                 
    Duration of severe neutropenia is critical to severe infection risk
    If the fever is prolonged—think fungus
    If the fever is recurrent—think fungus
    If there is pneumonia after fever—think fungus
    Fungal infection can occur even after resolution of neutropenia

Improved Diagnosis of Invasive Pulmonary Aspergillosis

The diagnosis of invasive pulmonary aspergillosis requires the identification of fungus by stain or culture in a patient with appropriate clinical risk factors and radiographic findings. The clinical settings include recent significant neutropenia and/or high-dose steroid use. Confirmation of tissue invasion is neither feasible nor required in most cases. Fever, although common among neutropenic patients, is not present in all patients with invasive pulmonary aspergillosis. The typical radiographic presentation is one or more focal infiltrates, usually densely consolidated because of lung infarction. Cavitation is a relatively late finding, and often is associated with recovery of the neutrophil count and resolution of the infection. CT scanning appears to be the most sensitive procedure to detect clinically significant invasive pulmonary aspergillosis.7

The diagnostic approach to invasive pulmonary aspergillosis often involves procedures to acquire respiratory samples for stain and culture. Stain provides a more rapid diagnosis than culture. It is unclear whether fungal culture or direct staining of samples (most often with methenamine silver) is more sensitive. The sensitivity likely depends on the source of the sample, the amount of material subjected to stain, and the number of stained slides examined.

The diagnostic yield of procedures to detect invasive pulmonary aspergillus varies. (See Table 3.) Bronchoscopy detects about half of cases with infection. Washings or bronchoalveolar lavage is more sensitive than biopsy because of the larger area of lung sampled. My personal experience suggests that open lung biopsy may miss the fungus in as many as 20% of lesions due to infection. I doubt that even autopsy reveals all the infections. It is disquieting to think that currently there may not be a gold standard for detection of invasive pulmonary aspergillosis. The absolute confirmation of invasive pulmonary aspergillosis may not be crucial. Among marrow transplant recipients there is no difference in outcome between those with "proven," as opposed to "suspected," invasive pulmonary aspergillosis.8 A strong clinical suspicion should to lead to treatment.

    Table 3
    Diagnostic Yield to Detect Invasive Pulmonary Aspergillus  
        Fine needle aspirate 50-67%
        Bronchoscopy ~50%
    (wash > BAL > transbronchial biopsy)
       Lung biopsy < 80%
       Autopsy < 100%??

A positive fungal culture or stain from the lower respiratory tract without overt evidence of invasive disease poses a dilemma. Several studies suggest that the incidence of invasive pulmonary aspergillosis among patients with neutropenia or high-dose steroid use is high (60-80%) after identification of Aspergillus in the airway. In general, it is advisable to assume that detection of Aspergillus in a high-risk patient represents true infection.9,10

The most promising technique to detect Aspergillus infection is the polymerase chain reaction (PCR). Einsele and colleagues reported the results of a multicenter study of fungal PCR using primers to a wide range of fungi, including Candida and Aspergillus species.11 The assay detected as little as 1 cfu/mL in blood. Einsele et al examined 601 blood samples from 35 controls and 86 patients. All the controls tested negative and all known infected patients tested positive. The PCR assay results preceded radiographic changes suggestive of invasive pulmonary aspergillosis by an average of four days in most cases (12 of 17). The assay became negative in patients who responded to antifungal therapy.11 Routine clinical application of such a technology will revolutionize the care of these patients. Much of the uncertainty of diagnosis and the empiricism of treatment will evaporate. I am excited for these assays to be perfected and anticipate they will be applied to the detection of fungal products in respiratory secretions as well as in blood.


Amphotericin B deoxycholate remains the standard treatment for invasive pulmonary aspergillosis. Unfortunately, the success rates remain poor. Many, if not most, of the patients with invasive pulmonary aspergillosis die of the underlying disease that placed them at risk for invasive pulmonary aspergillosis in the first place rather than of the fungal infection. In addition, recovery of an adequate and functional neutrophil count is the common feature among patients who recover from invasive pulmonary aspergillosis. This may explain why patients who are less intensely suppressed, such as renal or heart transplant recipients, seem to fare better than patients with leukemia and chemotherapy.12,13

Use of Amphotericin B is limited by toxicity, especially renal insufficiency. Recent interest in lipid-complexed Amphotericin B is based on the rationale that these highly lipophilic preparations have decreased toxicity to human cells due to a high affinity to fungal ergosterols. There are three commercially available preparations, and all are less toxic than conventional Amphotericin B. Amphocil (Amphotericin B Colloidal Dispersion or "ABCD") is a "ribbonlike" drug-phospholipid complex. Albecet (Amphotericin B Lipid Complex or "ABLC") is a "disk-like" complex of Amphotericin B and cholesteryl sulfate. AmBisome (Liposomal Amphotericin B) is the only true liposomal preparation and is composed of Amphotericin B and two phospholipids. While each has less toxicity than Amphotericin B, AmBisome displays the least and ABCD the most.14,15

The decreased toxicity profiles have led to the expectation that increased doses of the lipid-complexed Amphotericin B preparations could be administered. Unfortunately, the pharmacokinetics of these agents are difficult to follow. They are preferentially taken up by the reticuloendothelial system so that blood levels are of limited use. Exact dosing guidelines are uncertain. At this time, there are few data to suggest an improved response among patients receiving these agents.

Fluconazole is effective in Candida and Cryptococcal infections after organ transplantation. It has no activity against Aspergillus. Itraconazole, however, is an azole with activity against Aspergillus. It is only available in oral form in the United States. An intravenous preparation is under study. Unpredictable absorption limits its use in the acute care setting. Itraconazole may be a reasonable alternative in ambulatory settings with chronic forms of Aspergillus infection, but I believe it has little role in the treatment of invasive pulmonary aspergillosis in the ICU presently.

Invasive pulmonary aspergillosis often is a localized infection associated with substantial amounts of pulmonary infarction. Antifungal drugs are often ineffective. Therefore, I believe that surgical resection of infected tissue should be considered in patients with infection limited to the thorax who will likely remain neutropenic or on steroids. In addition, I consider resection in patients who appear to be at high risk of hemorrhage due to erosion into thoracic vascular structures. Successful resection of invasive pulmonary aspergillosis has been reported. However, it has been difficult to demonstrate a survival advantage because of competing causes of death in the patient populations affected.

There are alternative, and as yet unproven, strategies for treatment of invasive pulmonary aspergillosis. These include the use of granulocyte transfusions for neutropenic patients. Anecdotal reports exist, but there are no positive prospective studies. Similarly, the administration of hematopoietic growth factors to augment neutrophil number and function seems reasonable. Prophylactically administered growth factors in patients receiving treatment for lung cancer have had a favorable impact on infections in some studies. I believe there is sound rationale for their use in the neutropenic patient with invasive pulmonary aspergillosis.

The essential points in devising therapeutic strategies for invasive pulmonary aspergillosis are tempered by the recognition that it is associated with high mortality due to both the difficulty in treating the infection and the morbidity of the underlying conditions associated with the risks for infection.16 Persistent myelosuppresion is usually fatal and restoration of adequate functioning phagocytes is critical to recovery. More sensitive diagnostic studies are needed to detect and follow the infection. At this time, intravenous Amphotericin B remains the standard of care. Lipid-complexed Amphotericin should be reserved for those who fail or are intolerant. These agents are reasonable first-line approaches in the presence of pre-existing renal dysfunction. Lastly, I believe surgical resection should be considered in cases of limited thoracic infection. Table 4 summarizes these therapeutic strategies.

    Table 4 
    Therapeutic Strategies for Invasive Pulmonary Aspergillosis                             
    Invasive pulmonary aspergillosis associated with high mortality
    Persistent myelosuppresion is usually fatal
    More sensitive diagnostic studies needed
    Amphotericin B remains the standard of care
    Lipid-complexed drug should be reserved for those who fail or are intolerant
    Early surgical resection should be considered


Patients with invasive pulmonary aspergillosis are increasing in numbers in the ICU and we will need to become more familiar with this disease. It is a frustrating disease because many of the patients are profoundly immunosuppressed. The high mortality rate is in part related to the underlying disease, not necessarily the infection. I believe that vigilance is required because the presentation may be subtle and the infection unsuspected in some patients (such as those with COPD). More sensitive diagnostic modalities are on the horizon. In addition, I expect that more effective and less toxic therapeutic options will help us to help these patients.


    1. Pittet D, et al. Am J Respir Crit Care Med 1996; 154(2 Pt 1):541-544.

    2. Conesa D, et al. Ann Pharmacother 1995;29: 1235-1237.

    3. Kolbe K, et al. Bone Marrow Transplant 1997;19:143-147.

    4. Novakova IR, et al. Leuk Lymphoma 1993; 10:461-467.

    5. Maschmeyer G, et al. Cancer 1994;73:2296-2304.

    6. Barton TD, Schuster MG. Clin Infect Dis 1996; 22:1064-1068.

    7. Blum U, et al. Chest 1994;106(4):1156-1161.

    8. Wald A, et al. J Infect Dis 1997;175:1459-1466.

    9. Horvath J, Dummer S. Am J Med 1996;100:171-178.

    10. Levy H, et al. Respir Med 1992;86:243-248.

    11. Einsele H, et al. J Clin Microbiol 1997; 35:1353-1360.

    12. Biggs VJ, et al. Clin Infect Dis 1994;18:123-124.

    13. Wang SS, et al. Transplant Proc 1994;26:2329-2332.

    14. Hiemenz JW, Walsh TJ. Clin Infect Dis 1996; 22:S133-S144.

    15. Mills W, et al. Br J Haematol 1994;86:754-760.

    16. Denning DW. J Infect 1994;28 (Suppl1):25-33.