Fax: (713) 745-6839
Clinical significance of non-Candida fungal blood isolation in patients undergoing high-risk allogeneic hematopoietic stem cell transplantation (1993–2001)
Article first published online: 23 APR 2004
Copyright © 2004 American Cancer Society
Volume 100, Issue 11, pages 2456–2461, 1 June 2004
How to Cite
Safdar, A., Singhal, S. and Mehta, J. (2004), Clinical significance of non-Candida fungal blood isolation in patients undergoing high-risk allogeneic hematopoietic stem cell transplantation (1993–2001). Cancer, 100: 2456–2461. doi: 10.1002/cncr.20262
- Issue published online: 18 MAY 2004
- Article first published online: 23 APR 2004
- Manuscript Accepted: 10 MAR 2004
- Manuscript Revised: 25 FEB 2004
- Manuscript Received: 12 JAN 2004
- significant fungemia;
- filamentous fungi;
- hematopoietic stem cell transplantation
The clinical relevance of mold isolated from blood cultures, even in severely immunosuppressed allogeneic hematopoietic stem cell transplantation (HSCT) recipients, remains uncertain. The authors hypothesized that isolation of non-Candida fungi from blood cultures in patients undergoing high-risk HSCT would have clinical significance.
The authors reviewed the records of 73 allogeneic HSCT recipients between January 1, 1993 and January 1, 2001 in whom fungal species were isolated from blood cultures.
Fifty-two episodes of non-Candida fungemia occurred in 48 patients (66%) after a median of 10 days (range, 2–341) after transplantation. All 48 patients had indwelling intravascular catheters, and 23 patients (48%) had profound neutropenia. Thirty-five of 48 patients had received partially matched, related donor stem cell grafts (19 patients had 3-antigen-mismatched grafts); 35 patients had undergone T-cell depleted transplantation and 9 patients were receiving treatment for acute graft-versus-host disease. In 5 of 48 patients (10%), death was attributed to fungemia that occurred 8–11 days after the initial fungal blood culture was obtained; all 5 patients were age > 30 years. No deaths occurred in the younger age group (n = 22 patients; P = 0.05). In the 24 patients who did not receive systemic antifungal therapy, 4 deaths (17%) were attributed to infections with Penicillium (n = 2 patients), Epicoccum (n = 1 patient), or Penicillium plus Cladosporium species (n = 1 patient). Of the 24 patients who received amphotericin B, only 1 patient (4%) died as a result of a probable hematogenous Aspergillus species infection; this difference in outcome, however, was not significant (P = 0.2).
Most of the non-Candida fungal blood culture isolates in recipients of high-risk, mismatched donor transplantation were clinically nonsignificant. However, because these low-virulence saprophytes occasionally may cause life-threatening disease, a reevaluation of the existing diagnostic paradigm is needed so that clinically significant fungemia may be differentiated from pseudofungemia. Cancer 2004. © 2004 American Cancer Society.
Nonpathogenic, saprophytic fungal isolates from blood culture samples often are regarded as nondisease-causing environmental contaminants.1 The clinical significance of non-Candida fungal species in blood, even in profoundly immunosuppressed individuals, remains uncertain.2–4 However, although “clinically significant” or “definitive” fungemia due to environmental molds is uncommon,4 in the past 20 years these low-virulence microorganisms have emerged as important pathogens in patients with compromised immune systems, especially in recipients of allogeneic bone marrow and blood stem cell transplantation.5, 6 Of the various fungal organisms, the non-Aspergillus filamentous molds, such as Alternaria, Cladosporium, Curvularia, Fusarium, Paecilomyces lilacinus, Pseudallescheria boydii, Scedosporium apiospermum, and Scedosporium prolificans, often are not susceptible to commonly prescribed antifungals, including amphotericin B.5–10 Therefore, in patients with fungal infections that most likely will not respond well to amphotericin B, preemptive or empiric therapy with the new, broad-spectrum, triazole-based antifungals; the pneumocandin-echinocandin analogues; or both would be appropriate if clinical identifiers were available to distinguish “clinically significant” fungemia from pseudofungemia. Unfortunately, the current diagnostic guidelines are difficult to implement, because the required histologic analysis of the infected tissue seldom is available in these high-risk transplantation recipients because of the underlying severe thrombocytopenia. We performed this retrospective analysis to evaluate the clinical relevance of blood culture isolates of non-Candida fungal species in severely immunosuppressed patients undergoing mismatched allogeneic stem cell transplantation.
MATERIALS AND METHODS
A retrospective review was performed of all blood cultures positive for non-Candida species from recipients of stem cell transplantation between January 1, 1993 and January 1, 2001 at Palmetto-Richland Memorial Hospital Cancer Center (Columbia, SC). Patient and laboratory information was retrieved from patients' charts and the computerized hospital data systems. All values are given in medians and ranges.
All blood culture samples were processed by the automated BACTEC 9240 system (Becton Dickinson Microbiology System, Sparks, MD) at the hospital's microbiology laboratory. The blood cultures were incubated for 5 days in aerobic and anaerobic culture media: BACTEC PLUS and LYTIC10, respectively. Fungal species were identified using standard methods.11 The South Carolina Department of Health and Environmental Control mycology laboratories in Columbia served as an extrainstitutional reference for the nonidentifiable microorganisms.
Blood cultures were obtained from all transplantation recipients who had fever, clinical signs of infection, or both; surveillance blood cultures, however, were not obtained routinely. Probable fungemia was defined as a clinical illness associated with and consistent with the presence of non-Candida fungal species in blood culture specimens. Patients with histologic evidence of invasive fungal disease were considered to have “definite” or “true” fungemia.4 Pseudofungemia was defined as the presence of a non-Candida fungal species isolated from the blood culture of a patient with a clinical illness due to another known pathogen, with features inconsistent with systemic mycoses, or with both.3, 4 All positive blood cultures obtained during the same hospitalization were considered a single episode; if more than one species was isolated, then each was regarded as a separate episode. Attributable mortality was defined as death within 30 days of the last positive fungal blood culture and in the absence of a known terminal event, such as intracranial or severe gastrointestinal hemorrhage, pulmonary embolism, myocardial infarction, pneumonitis due to human cytomegalovirus, bacterial septic shock, adult respiratory distress syndrome, or progressive graft-versus-host disease (GVHD). The terms granulocytopenia and neutropenia are used interchangeably herein, and profound neutropenia was considered to be an absolute neutrophil count (ANC) < 100 cells/μL.
Conditioning Regimen and GVHD Prophylaxis
Antifungal Therapy and Intravascular Catheter Removal
All clinically ill patients with hematogenous mycoses received treatment with systemic, polyene-based antifungals, including amphotericin B deoxycholate (0.7–1.0 mg/kg daily) and liposomal amphotericin B compound (3.0–5.0 mg/kg daily). Indwelling central venous catheters were removed immediately from severely ill patients and from those receiving high-dose corticosteroids, antithymocyte globulin for prophylaxis or treatment of GVHD, or both agents. In patients who did not have their catheters removed, the antifungal drug was administered through all catheter lumens. All intravascular lines (central and peripheral) were removed from patients with persistent (≥ 72 hours) positive blood cultures.
Chi-square and Fisher exact tests were used to compare categoric variables. The differences between these variables were considered significant if the two-tailed P value was ≤ 0.05.
Patient information is given in Table 1. During the 8 years under review, 52 episodes of non-Candida fungemia occurred after a median of 10 days (range, 2–341 days) in 48 patients after transplantation. The frequency of non-Candida species fungal blood culture isolation was 9 in 100 transplantations (35 of 379 recipients) in the recipients of partially matched-related donor stem cell grafts; 15 in 100 transplantations (5 of 34 recipients) in recipients of matched unrelated donor grafts; 5 in 100 transplantations (3 of 61 recipients) in recipients of matched related donor grafts; and 4 in 100 transplantations (5 of 115 recipients) in recipients of autologous stem cell grafts (Table 2).
|Characteristic||No. of patients (%)|
|No. of patients||48|
|Median age ± SD (yrs)||40 ± 17|
|Absolute neutrophil count ≤ 100 cells/μL||23 (48)|
|Underlying malignancy||48 (100)|
|Acute myelogenous leukemia||10 (21)|
|Acute lymphocytic leukemia||9 (19)|
|Chronic myelogenous leukemia||11 (23)|
|Chronic lymphocytic leukemia||1 (2)|
|Multiple myeloma||2 (4)|
|Myelodysplastic syndrome||4 (8)|
|Non-Hodgkin lymphoma||5 (10)|
|Hodgkin's lymphoma||2 (4)|
|Stem cell grafts||48 (100)|
|Partially matched related donors||35 (73)|
|One HLA mismatch||1 (3)|
|Two HLA mismatches||15 (43)|
|Three HLA mismatches||19 (54)|
|Matched related donors||3 (6)|
|Matched unrelated donors||5 (10)|
|T-cell depleted grafts||35 (73)|
|Acute graft-versus-host disease||9 (19)|
|Patient||Age (yrs)||Gender||Transplantation (antigen mismatches)||Cancer||T-cell depletion||Steroids||Acute GVHD||Days post HSCT||Leukocyte count (× 109L)||Fungal species (no. of cultures)||Antifungal prophylaxis||Antifungal therapy||Died after HSCT (days)|
|5||4||Female||PMRD (2)||ALL||+||+||−||24||1800||Trichosporon beigelii, Aureobasidium||None||AMB||—|
|7||7||Male||PMRD (2)||AML||+||+||−||21||1600||Trichosporon beigelii||None||AMB||—|
|8||8||Male||Autologous||NB||−||−||−||8||< 100||A. flavus||None||AMB||—|
|9||8||Female||PMRD (3)||ALL||+||+||+||17||2700||Cladosporium (2)||AMB||None||—|
|12||13||Male||PMRD (2)||ALL||+||+||−||222||200||A. niger, Penicillium||Fluconazole||None||—|
|18||22||Female||PMRD (2)||AML||+||+||−||4||< 100||Geotrichum||Fluconazole||None||—|
|19||23||Male||PMRD (3)||MDS||+||+||+||10||< 100||Penicillium||Fluconazole||None||—|
|21||26||Male||PMRD (3)||HD||+||+||−||20||2400||Geotrichum (2)||Fluconazole||None||—|
|24||31||Male||PMRD (2)||CML||+||+||−||9||< 100||Aureobasidium||Fluconazole||AMB||—|
|26||31||Male||PMRD (3)||ALL||+||+||−||48||13,900||Cladosporium, Penicillium||None||None||—|
|27||32||Male||MRD||CML||−||−||−||12||< 100||Aspergillus spp.||None||Abelcet||23|
|29||35||Male||PMRD (3)||CLL||+||+||−||46||14,600||Aspergillus spp.||None||Abelcet||—|
|33||39||Male||PMRD (2)||CML||+||+||−||4||< 100||Cladosporium||Fluconazole||None||—|
|36||44||Male||PMRD (2)||AML||+||+||+||6||< 100||A. fumigatus (2), Alternaria||AMB||None||—|
|37||44||Male||PMRD (3)||MM||+||+||+||34||4600||Penicillium (2)||AMB||AMB||—|
|38||44||Female||PMRD (3)||MDS||+||+||−||7||< 100||Penicillium||Fluconazole||None||—|
|39||46||Female||PMRD (3)||AML||+||+||−||5||< 100||Aspergillus spp.||None||AMB||—|
|42||49||Male||PMRD (3)||AML||+||+||−||11||< 100||Penicillium (3)||Abelcet®||None||—|
|44||52||Male||PMRD (2)||MDS||+||+||−||5||< 100||Penicillium||Fluconazole||None||13|
|47||58||Male||PMRD (2)||AML||+||+||−||5||< 100||Penicillium||Fluconazole||None||—|
The fungal species identified were Penicillium (20 episodes), Cladosporium (7 episodes), Aspergillus species (A. fumigatus) (1 episode), A. flavus (1 episode), A. niger (1 episode) and 3 other Aspergillus species, Geotrichum candidum (5 episodes), Chrysosporium (3 episodes), Trichosporon beigelii (2 episodes), and Aureobasidium (2 episodes). Single episodes due to Alternaria, Curvularia, Epicoccum, Paecilomyces, Scopulariopsis, Sporobolomyces, and Saccharomyces cerevisiae also were observed. All but five patients had two or more blood cultures that were positive for a mold (A. fumigatus, Cladosporium, Geotrichum, and Penicillium species). In 2 profoundly neutropenic patients (ANC < 100 cells/μL), no deaths were observed despite the presence of multiple positive fungal blood isolates.
Twenty-seven patients had 29 episodes of Candida species bloodstream invasion during 8 years. Candida krusei was the most common (n = 11 patients; 38%), followed by C. glabrata (n = 9 patients; 31%), C. albicans (n = 4 patients; 14%), C. tropicalis (n = 3 patients; 10%), and C. parapsilosis (n = 2 patients; 7%). Two patients had concomitant non-Candida species fungemia (Patients 12 and 29) (Table 2).
In 5 of 48 patients (10%), deaths were attributed to fungemia that occurred due to hematogenous mycosis 8–11 days after the initial fungal blood isolation; 3 of those 5 patients (60%) had profound granulocytopenia (ANC < 100 cells/μL) (Table 2). All deaths occurred in patients age > 30 years compared with no deaths in younger patients (P = 0.05). Among the 24 patients who did not receive systemic antifungal therapy, 4 patients (17%) died of probable hematogenous mycosis due to Penicillium (n = 2 patients), Epicoccum (n = 1 patient), or Penicillium plus Cladosporium species (n = 1 patient). In contrast, among the patients who were given amphotericin B (n = 24 patients), only 1 death (4%) was attributed to “probable” aspergillemia; this disparity in outcome, however, was not found to be significant (P = 0.2) (Table 2). Two of 13 neutropenic patients (15%) who did not receive amphotericin B died, whereas 2 of 8 untreated patients (25%) with normal neutrophil counts died.
Most of the non-Candida fungal blood culture isolates in our transplantation patients (with a predominant population of severely immunosuppressed recipients of two to three antigen-mismatched stem cell grafts) represented clinically nonsignificant fungemia or pseudofungemia. In concert with our observation, aspergillemia in patients with cancer who did not undergo hematopoietic stem cell transplantation seldom was observed as “clinically significant” or “definite” hematogenous mycosis.3, 4 Similar to the current study (Table 2), other studies have shown that the presence of non-Candida fungal blood isolates in most transplantation recipients with profound immunosuppression most likely represents contamination from an environmental source, suggesting “clinically nonsignificant fungemia” or “pseudofungemia.”1, 2 In a retrospective analysis of Aspergillus species fungemia in patients with cancer, nearly 20% of all Aspergillus species-positive blood cultures were reported as representing “definite” mycoses; however, using the existing guidelines, 86% of those diagnoses could be established on postmortem examination only.3 This inadequacy of positive and negative predictive values of the existing diagnostic tests raises serious concerns regarding the limitations of the existing parameters for diagnosing “clinically significant” bloodstream invasion due to low-virulence, saprophytic, environmental molds.2, 4
Patients with hematologic malignancies,3, 14 recipients of hematopoietic stem cell transplantation, and patients receiving cytotoxic therapy for posttransplantation complications, such as GVHD, are at increased risk for disseminated fungal infections of non-Candida organisms.15–17 Traditionally, most systemic fungal infections have been observed during the early allogeneic stem cell transplantation period (< 180 days after the transplantation), although, in patients with prolonged, severe immune dysfunction due to chronic GVHD therapy, late reactivation of human cytomegalovirus, etc., invasive mycoses may be encountered > 360 days after transplantation.18 Because of the risk of serious hemorrhage in patients with low platelet counts (< 50,000 cells/μL), histologic examinations of the involved or infected organ are often deferred3, 4, 19; this problem is highlighted by the paucity of antemortem diagnoses of definite invasive mycoses and the fact that most definite invasive fungal infections diagnosed with current criteria are reported in patients only after death.3, 10, 20 Fortunately, in the current study, the fungemia-associated attributable mortality (15%) was found to be low among neutropenic allogeneic stem cell transplantation recipients who did not receive antifungal therapy and suggests that most fungal blood isolates (85%), even in a highly susceptible group, were “clinically nonsignificant” and most likely represented environmental contamination (Table 2).2–4
Diagnostic parameters for “probable” fungemia are arbitrary; therefore, surrogate diagnostic markers21 are needed critically to reduce the variability in interpretations of clinical data and the inconsistencies in decisions to either withhold or institute effective therapy. The current criteria, which are far from satisfactory, suggest probable systemic mycoses; they include 1) an immunologically susceptible host (severe granulocytopenia, cellular immune dysfunction, or both), 2) breakthrough fungal infections in patients receiving subtherapeutic antifungals as prophylaxis, and 3) radiographic and clinical features consistent with fungal tissue invasion and tissue damage.2, 18 Detection of fungal antigens, such as galactomannan or D-glucan, in patients' sera may provide important adjuvant information regarding the source of Aspergillus or non-Aspergillus filamentous mold in blood culture samples22–25; however, nondetectable antigenemia in patients receiving prophylactic or empiric antifungal therapy, such as itraconazole and amphotericin B, may cause serum antigen levels to go undetected and to be misconstrued as (false-) negative results.26 Serial detection of fungus-specific molecular markers by nucleic acid hybridization and amplification techniques in the host's blood, in bronchioalveolar lavage sample, or in samples from other infected sites may serve as an important adjuvant tool in determining clinically significant fungemia.27 These methods also may improve the ability of clinicians to make informed, evidence-based decisions when considering treatment options for immunosuppressed cancer patients or stem cell transplantation recipients with positive blood cultures for a non-Candida fungal species.28, 29 Detection of fungal cell wall components22 as well as organism-specific and class-specific nucleic acid amplification assays30 appears promising in diagnosis of clinically significant fungemia and pseudofungemia. We recognize that a critical evaluation of the new and old diagnostic laboratory tests is needed to establish a comprehensive protocol that can be applied with relative ease in hospitalized patients to improve the accuracy of antemortem diagnoses of invasive fungal infections.
In the current study, clinically significant fungemia due to non-Candida species was uncommon in immunosuppressed stem cell transplantation recipients. All the deaths reported in the older patients suggested a greater probability of definite fungal dissemination in transplantation recipients with advancing age. The existing clinical parameters for probable non-Candida fungemia are arbitrary, and the new molecular surrogate diagnostic markers should be evaluated further to enhance the capability of clinicians to diagnose clinically significant, invasive fungal infections, especially in patients with hematologic malignancies and in recipients of hematopoietic stem cell transplantation.
The authors are especially grateful to Dr. Jean P. Henslee-Downey and Shannon DeRienzo of the University of South Carolina School of Medicine for their support in this project.
- 1Bloodstream infections: laboratory detection and clinical considerations. Chicago: American Society of Clinical Pathology Press, 1998., .
- 7New spectrum of fungal infections in patients with cancer. Rev Infect Dis. 1989; 1: 369–378., , , et al.
- 10Breakthrough Scedosporium apiospermum (Pseudallescheria boydii) brain abscess during therapy for invasive pulmonary aspergillosis following high-risk allogeneic hematopoietic stem cell transplantation: Scedosporiasis and recent advances in antifungal therapy. Transplant Infect Dis. 2002; 4: 212–217., , .
- 11Manual of clinical microbiology. Washington, DC: American Society for Microbiology, 1995., , , , .
- 18Infections in patients with neoplastic diseases. In: ShoemakerWC, GrenvikM, AyersSM, HolbrookPR, editors. Textbook of critical care. Philadelphia: WB Saunders, 2000: 715–726., .
- 25Prospects for the early diagnosis of invasive aspergillosis in the immunocompromised patients. Rev Med Microbiol. 1996; 7: 105–113., , , .