Fungal infections after bone marrow transplant

Fungal infections after bone marrow transplant

What every physician needs to know:

Invasive fungal infections

Infections caused by yeasts, especially Candida species, used to be common after bone marrow transplant (BMT), before the widespread use of antifungal prophylaxis.

Even with azole prophylaxis, Candida species can break through as a cause of systemic disease, manifesting most commonly as either a bloodstream infection (candidemia), or as isolated organ disease (hepatosplenic candidiasis). Candida rarely cause lung infections, which more frequently involve filamentous fungi that are inhaled from an environmental source, such as Aspergillus species, endemic fungi (for example, histoplasmosis, cryptococcosis), or Pneumocystis.

Aspergillus species are the most common cause of pulmonary fungal infections, and these infections can occur during primary periods of neutropenia and graft-versus-host disease (GVHD). Discussion here is focused on candidiasis and pulmonary fungal infections.

What features of the presentation will guide me toward possible causes and next treatment steps:

Prevention of candidal infections is key, and has been shown to improve survival in the highest risk patients. Hence, when a Candida infection is diagnosed in a patient who is receiving azole prophylaxis, it is usually caused by one of the azole-resistant species, such as Candida glabrata or Candida krusei. This limits the use of azole antifungals therapeutically.

Candida infections should be considered as a cause of infection in any BMT recipient who has therapy or GVHD-induced mucositis, and/or in the presence of an indwelling intravascular catheter. Most frequently, candidemia presents as a febrile illness, potentially with sepsis-like manifestations. Skin rashes can occur, and appear most frequently as erythemetous papules or pustules.

Fungi, especially filamentous organisms such as Aspergillus species, should be considered with presentation of focal pulmonary infiltrates, such as nodules, with or without "halo" or cavitation. In BMT recipients, invasive aspergillosis can present with varied radiographic abnormalities, ranging from isolated to multiple nodules to larger focal infiltrates, sometimes at the same time (Figure 1). These organisms disseminate through the bloodstream, most frequently causing skin lesions and central nervous system lesions.

Figure 1.

Invasive aspergillosis can present with varied radiographic abnormalities, ranging from isolated to multiple nodules to larger focal infiltrates, sometimes at the same time.


Fungi cause both invasive and allergic sinusitis; in the BMT recipient, invasive sinusitis with dissemination into the orbit and/or brain is the most severe manifestation. Biopsy of the sinuses is important both to confirm the diagnosis of invasive fungal sinusitis and to identify which organism is the cause, as therapy differs depending on the organism.

What laboratory studies should you order to help make the diagnosis and how should you interpret the results?

Culture of blood, although not sensitive enough to rule-out Candida as a cause of a febrile syndrome, is important to guide therapy, and to measure success of antifungal treatment. Blood cultures should be performed daily after a patient develops candidemia, until a negative result is obtained; in order to minimize the likelihood of organ involvement, therapy should be administered for at least 10 days after the last positive blood culture.

Candida species can present as hepatosplenic candidiasis, especially in people who have had prior episodes of neutropenia concurrent with gastrointestinal (GI) mucositis; risks are especially high in people who have received such therapies in the absence of prophylactic antifungals. This diagnosis should be suspected in people who present with elevated liver enzymes, especially alkaline phosphatase, and fever, potentially with abdominal or flank pain. In this setting, multiple hypodense lesions are usually apparent on ultrasound, computed tomography (CT) scan or magnetic resonance imaging (MRI).

While liver biopsy may be useful to document the cause of disease, and to distinguish between yeasts and other causes of liver abscesses (such as bacteria), negative findings do not rule out hepatosplenic candidiasis, as frequently, the inciting yeast is difficult to find, and the main histopathological finding points to the resultant (frequently granulomatous) inflammatory response.

If a patient presents with pulmonary abnormalities (nodular opacities or focal infiltrates), that are suggestive of an invasive fungal infection, one should obtain a serum galactomannan assay, and consider bronchoscopy. A positive galactomannan test is highly predictive of invasive aspergillosis when found in the presence of nodular or focal infiltrates.

Bronchoalveolar lavage (BAL) is an essential first step to diagnosing pulmonary fungal infections, and several studies have demonstrated safety. Use of the galactomannan assay on BAL fluid adds to the sensitivity of diagnosing invasive aspergillosis, and it should be done in addition to culture.

Utility of culture

Any blood culture that reveals Candida species should be treated as a true candidemia, and not attributed to colonization or contamination; one of the reasons that we treat all positive blood cultures, is that there is an increased risk for subsequent embolic complications, such as endophthalmitis, in people who have "transient" candidemia.

Blood cultures that reveal filamentous organisms are usually contaminants from the laboratory, except with growth of one of the fungi that can sporulate in vivo, such as Fusarium species.

Utility of antigen-based assays

It is important to realize that blood cultures are not very sensitive for Candida species. While newer technologies may enrich for Candida growth, these organisms should be considered as potential causes of febrile illnesses, even in the absence of positive cultures. Several tests that detect circulating beta-D-glucan in blood are being used in medical centers throughout the world. While these assays are sensitive for detection of fungal components, results are positive in the presence of multiple fungi, including filamentous organisms that cause pulmonary disease. Also, there are multiple causes of false-positive test results, as beta-glucan can be found in many medical products (for example, gauze, dialysis filters).

Galactomannan is an antigen that is expressed by Aspergillus species, and some organisms that are closely related (such as Penicillium species. This has become an important diagnostic test, applied to both blood and BAL fluid. False-positive tests can occur, especially in people who are receiving drugs that are contaminated with galactomannan, such as piperacillin-tazobactam.

What conditions can underlie fungal infections after bone marrow transplant

What conditions can underlie invasive fungal infections?

People who are receiving therapy for hematologic malignancies have multiple risks for candidiasis, largely due to poor immunologic defenses and breakdown of skin and gastrointestinal (GI) tract barriers. Fungal infections involving the lung occur due to poor neutrophilic and/or cellular immunity, after inhaled exposure of the organisms.

When do you need to get more aggressive tests:

More aggressive tests to document the microbial cause of any systemic illness should be considered when signs and symptoms of the infection are not responding to what should be appropriate antifungal therapy. For instance, biopsy of skin is a good idea when a rash is expanding on antifungal therapy, and biopsy of the liver may be necessary to define the microbial etiology of an expanding liver lesion. However, the caveat is that liver lesions are anticipated to grow even with appropriate antifungal therapy, as this is usually caused by inflammation itself.

When a pulmonary lesion is not responding to an empirical antifungal course, biopsy by video-assisted thoracoscopic surgery (VATS)should be considered. I believe that this is very important today, when antifungal therapy with voriconazole may only target Aspergillus species, and not include other filamentous organisms, such as agents of mucormycosis, or bacterial pathogens. It is also important in ruling out non-infectious causes of pulmonary lesions, such as bronchiolitis obliterans.

Skin lesions

The presence of multiple skin lesions in a patient who has pulmonary findings consistent with fungal pneumonia are suggestive of hematogenous spread, especially with organisms that can undergo in vivo sporulation, such as Fusarium species, Alternaria species, and Scedosporium species. Biopsy should be performed to document microbial etiology, as these infections need to be very aggressively treated and prognosis is poor.


Invasive fungal sinusitis cannot be ruled out by a "normal" visual examination; biopsy, even blind biopsies performed in the absence of visual necrosis or lesions are necessary to identify microscopic hyphal mucosal involvement.

What imaging studies (if any) will be helpful?

It is not usually necessary to perform additional imaging studies in the presence of a positive blood culture, except in the setting of suspected peripheral manifestations, such as endocarditis (rare, evaluate with echocardiography), or with suspected intra-abdominal or renal abscesses. MRI is more sensitive compared to either ultrasound or CT scan in detecting hepatosplenic candidiasis.

Patients with invasive aspergillosis do not need routine CT scans of the head unless there are neurological findings that suggest dissemination.

What therapies should you initiate immediately and under what circumstances – even if root cause is unidentified?


In the setting of blood culture revealing an unidentified yeast, I usually initiate therapy with an echinocandin antifungal (caspofungin, micafungin, or anidulafungin). These drugs have good activity against the most common causes of invasive infection in people receiving azole prophylaxis, except for some species, such as Candida parapsilosis. If the organism is identified as one of these species that may demonstrate relative resistance to echinocandins, therapy is changed to a lipid amphotericin formulation. I usually do not initiate therapy with the lipid formulations of amphotericin, although these are acceptable treatments of candidemia, given toxicity profiles.

Hepatosplenic candidiasis

Suspect Candida as a cause of hypodense liver lesions, especially in a patient who is relative azole naive and has had a significant period of neutropenia and mucositis; in these patients, antifungal therapy that targets Candida albicans, such as an echinocandin, is the appropriate first step, even without documenting infection with liver biopsy.

Suspected invasive aspergillosis

Therapy for pulmonary fungal infections should be targeted to the organism that is found on culture of lavage fluid (or other respiratory samples). However, in the absence of culture confirmation, I usually start voriconazole (dosed according to weight) as therapy, except in patients who have had a high amount of exposure in the past, or those who have liver dysfunction or drug interactions that prohibits its use. In the latter situation, a lipid formulation of amphotericin is appropriate.

Documented mucormycosis

I administer relatively high doses of lipid formulations of amphotericin B (for example, 5 to 7.5mg/kg liposomal amphotericin B) with recovery of an agent of mucormycosis from biopsy or lavage.

What other therapies are helpful for reducing complications?

The utility of combination antifungal therapy is controversial and not warranted in the setting of candidemia. Intravascular catheters should be removed when possible, even in people with suspected invasion through the GI tract, as the catheter tip can be a nidus for secondary seeding and persistent infection. Although not demonstrated to have a documented impact in randomized trials, people with persistent candidemia and/or sepsis may benefit from granulocyte transfusions, especially if prolonged neutropenia is anticipated.

I always consider adding an echinocandin to a voriconazole or an amphotericin formulation in the setting of severe pulmonary invasive aspergillosis. A randomized trial that was recently did not definitively show that the combination of voriconazole and anidulafungin were superior to voriconazole alone, but there was a strong trend to improved survival at 6 weeks after invasive aspergillosis diagnosis in the people who received combination antifungal therapy; I believe that the risk-benefit ratio supports combined agents for primary therapy of documented disease.

Use of granulocyte transfusions in the setting of pulmonary fungal infections is not well supported by data, and considerable risk for pulmonary toxicities decrease my enthusiasm for using them.

Surgical resection of pulmonary fungal lesions should be considered when infiltrates are closely approximating large vessels, and in the setting of severe hemoptysis. Large cavitary lesions do not necessarily have to be resected, and can regress in size with appropriate therapy. Pulmonary fungal infections caused by Zygomycetes should be surgically debrided as there is typically a good amount of fungus within tissue necrosis.

Treatment of mucormycosis

High doses of a lipid formulation of amphotericin B are a gold-standard treatment. There are some in vitro and in vivo data suggesting that combination therapy with an echinocandin may help; I usually do add these (despite randomized controlled trials) in the setting of severe disease. Posaconazole is an azole that has good activity against these organisms, and it can be used after an initial "induction" regimen with the amphotericin formulation. Surgical resection of large necrotic lesions is a must for these organisms, which can grow even under necrotic conditions.

Empirical antifungal therapy

Many drugs, azoles, echinocandins, or amphotericin B formulations, have been studied and can be administered in the setting of suspected, but not culture documented candidiasis. However, prior azole exposure usually limits the utility of azole antifungals in BMT recipients.

What should you tell the patient and the family about prognosis?

Attributable mortality of candidiasis is relatively low and difficult to estimate, given the complex medical conditions that exist at time of candidemia. Most people clear Candida bloodstream infections without peripheral complications or death, however, studies have shown that approximately 20 to 30% of people who develop this infection die within 30 days (most likely due to multiple underlying causes, and sometimes, sepsis).

Death after diagnosis of aspergillosis is relatively common, occurring in approximately 30 to 50% of people within 12 weeks of diagnosis in the most recent studies, although the actual attributable mortality is difficult to estimate (likely approximating 20%).

“What if” scenarios.

Endophthalmitis is one of the more frequent peripheral manifestations of Candida species bloodstream invasion. Eye exams should be performed on people with positive blood cultures. If there are signs of endophthalmitis or vitritis, vitrectomy may be necessary. As azole antifungals penetrate the eye best, I usually add an azole (fluconazole, unless the bloodstream organism is C. krusei, in which case, voriconazole has good activity).

It is important to dose voriconazole aggressively, as people who are receiving the oral formulation at 200mg twice daily, frequently do not have high enough levels. Therapeutic drug monitoring by high-performance liquid chromatograpy (HPLC) is likely a good idea, if results are available in reasonable time.


Candida species exist as normal commensals in the GI tract and on the skin. Most invasive infections in BMT recipients occur through GI tract invasion, during period of mucositis and neutropenia. This is especially likely in someone who is not receiving azole antifungals, as C. albicans, the most pathogenic of the Candida species (and also very azole sensitive), can form hyphae to invade through the gut wall. Dissemination can thus occur via invasion into the systemic circulation or portal vasculature.

Peripheral manifestations can develop, most frequently involving the eye, heart valves, or skin. Invasion into the portal vasculature can lead directly to hepatic candidiasis, which may not present at time of infection, but rather, after neutrophils return, and evoke chronic inflammation. This is the reason that hepatosplenic disease can result in patients who never have any documented episodes of candidemia.

Filamentous fungi are ubiquitous in the environment. Spores (conidia) are inhaled frequently into the airways. In the absence of an effective innate immune response, conidia sporulate into germinated (hyphal) forms, which can both invade the lung tissue and cause inflammation. The filamentous forms are angion invasive and can disseminate through the bloodstream to distant sites (skin, brain, liver, et cetera).

What other clinical manifestations may help me to diagnose fungal infections after bone marrow transplant?

Colonization with Candida species predicts invasive disease, especially through the GI tract. The same is true for anyone noted to be colonized with Aspergillus species, especially in the respiratory tract. The exception is Aspergillus niger, which is a frequent inhabitant of the GI tract, and not usually associated with higher risks for infection.

What other additional laboratory studies may be ordered?


What’s the evidence?

Pappas, PG, Kauffman, CA, Andes, D. "Infectious Diseases Society of America. Clinical practice guidelines for the management of candidiasis: 2009 update by the Infectious Diseases Society of America". Clin Infect Dis. vol. 48. 2009. pp. 503-35.

[Comprehensive guidelines for treatment of candidal infections.]

Marr, KA, Seidel, K, White, TC, Bowden, RA. " Candidemia in allogeneic blood and marrow transplant recipients: evolution of risk factors after the adoption of prophylactic fluconazole". J Infect Dis. vol. 181. 2000. pp. 309-16.

[A good study that demonstrates microbial epidemiology and patient risks.]

Segal, BH. "Aspergillosis". New Eng J Med. vol. 360. 2009. pp. 1870-84.

[A good review focused on current understanding of pathophysiology, presentation, and therapy.]
You must be a registered member of Clinical Advisor to post a comment.

Sign Up for Free e-newsletters