Risk factors for invasive aspergillosis (IA) are incompletely identified and may undergo changes due to differences in medical practice. A cohort of 189 consecutive, adult patients with neutropenia hospitalized in the hemato-oncology ward of the University hospital Berne between 1995 and 1999 were included in a retrospective study to assess risk factors for IA. In total, 45 IA cases (nine proven, three probable, 33 possible), 11 patients with refractory fever and 133 controls were analyzed. IA cases had more often acute leukemia or myelodysplastic syndrome (MDS) (88 vs 38%, P<0.001) and a longer duration of neutropenia (mean 20.6 vs 9.9 days, P<0.001). They also had fewer neutropenic episodes during the preceding 6 months (mean 0.42 vs 1.03, P<0.001), that is, confirmed (82%) and probable (73%) IA occurred most often during the induction cycle. A short time interval (⩽14 days) between neutropenic episodes increased the risk of IA four-fold (P=0.06). Bacteremia, however, was not related to the number of preceding neutropenic episodes. Therefore, neutropenic patients with leukemia or MDS have the highest risk of IA. The risk is highest during the first induction cycle of treatment and increases with short-time intervals between treatment cycles.
Invasive infections due to Aspergillus species, mainly Aspergillus fumigatus, are an important cause of morbidity and mortality in patients with hematologic malignancies.1 The incidence of invasive aspergillosis (IA) may be as high as 22%2 and the lethality reaches 60–70% when IA occurs during neutropenia.3 An aggressive diagnostic approach and prompt treatment with antifungals are essential for survival in patients at risk for IA.4
Several risk factors for IA in patients with hematological malignancies have been identified.5 Probably the most important determinant is prolonged neutropenia.6 Profound neutropenia of more than 21 days duration increases the risk of IA four-fold. Administration of steroids,7, 8, 9 the type of chemotherapy,10 and qualitative disorders of granulocyte function11 have also been shown to increase the risk of IA. However, changes in the population at risk for IA due to intensified and new treatment strategies require a continuous reevaluation of risk factors for IA.
At present, prevention of IA concentrates mainly on measures that aim at reducing the risk of exposure to aspergillus spores, such as special ventilation of patient's rooms with HEPA filters and laminar flow, restriction of food, and isolation measures during construction activities in the hospital.12 Prophylaxis with antifungals may be beneficial, and may become a standard in the future.12 To optimally utilize prophylactic measures, careful characterization of patient groups at highest risk of IA is important. This may also facilitate earlier diagnosis and initiation of antifungal therapy.7
We conducted a retrospective study at the medical hemato-oncology ward in order to identify risk factors for IA among neutropenic patients with hematologic malignancies.
The medical oncology ward
The medical hemato-oncology ward at the University hospital Berne comprises nine beds for patients receiving intensive chemotherapy with or without autologous stem cell transplantation. Most patients cared for in this unit have hematological malignancies, including acute leukemias, myelodysplastic syndrome (MDS), aggressive lymphomas, multiple myelomas and solid tumors, and receive standardized treatment within prospective clinical trials. Patients with de novo AML were enrolled in a prospective trial and randomly assigned to receive G-CSF or no G-CSF during remission-induction cycles 1 and 2. Cycle 1 consisted of cytarabine (200 mg/m2 given by continuous infusion on days 1–7) and idarubicin (12 mg/m2 i.v. on days 6–8). Cycle 2 consisted of cytarabine (1000 mg/m2 i.v. over a period of 2 h every 12 h on days 1–6) and amsacrin (120 mg/m2 i.v. over a 60-min period on days 4–6). Patients who were in complete remission after cycle 2 were randomly assigned to a third cycle of chemotherapy with etoposide (100 mg/m2 i.v. on days 1–5) and mitoxantrone (10 mg/m2 i.v. on days 1–5) or high-dose chemotherapy with busulfan (4 mg/kg body weight p.o. daily on days 1–4) and cyclophosphamid (60 mg/kg body weight i.v. daily on days 5 and 6) followed by autologous stem-cell transplantation. Patients with MDS were treated with two cycles of remission-induction. Cycle 1 consisted of idarubicin (12 mg/m2 i.v. on days 1, 3, and 5), etoposide (100 mg/m2 i.v. on days 1–5) and cytarabine (100 mg/m2 given by continuous infusion on days 1–10). Cycle 2 consisted of intermediate-dose cytarabine (1000 mg/m2 given i.v. over a period of 2 h daily on days 1–6). Patients with relapsed AML received two cycles of high dose cytarabin (3000 mg/m2 given i.v. over a period of 3 h every 12 h on days 1–4) and mitoxantrone (12 mg/m2 i.v. on days 3–5). Patients with de novo ALL have received a first cycle with idarubicin (9 mg/m2 i.v. on days 1, 2, 3, and 8), cyclophosphamid (750 mg/m2 on days 1 and 8), vincristine (2 mg i.v. on days 1, 8, 15 and 22) and prednisolone (60 mg/m2 p.o. on days 1–7 and 15–22). Cycle 2 consisted of intermediate dose cytarabine (1000 mg/m2 given i.v. every 12 h on days 1–4) and mitoxantrone (10 mg/m2 i.v. on days 3–5). The conditioning regimens prior to autologous stem cell transplantation were mitoxantrone/melphalan (mitoxantrone 60 mg/m2 i.v. on day 1 and melphalan 180 mg/m2 on day 4) or BEAM (BiCNU 300 mg/m2 i.v. on day 1, melphalan 180 mg/m2 on day 1, etoposide 150 mg/m2 i.v. every 12 h on days 1–4 and cytarabine 200 mg/m2 i.v. every 12 h on days 1–4) for lymphomas, and high-dose melaphalan (200 mg/m2) for multiple myelomas. Patients with AML, MDS, lymphoma or multiple myeloma received methylprednisolon 125 mg per day as an additive antiemetic prophylaxis during chemotherapy days.
A specialist team of nurses, hematologists, oncologists and infectious disease specialists are responsible for the treatment and supportive care. During the study period, patients were hospitalized in single-isolation rooms with fore-chamber and positive air pressure ventilation. Additional hygiene measures were taken, when patients became neutropenic (neutrophil count <500/μl and/or leukocyte count <109/l). Patients left their room only for diagnostic procedures and had to wear a surgical mask outside their room. Special care was taken for routine disinfection of body orifices. Neutropenic patients with a new fever (>38.0°C for >1 h, >38.3°C, or shaking chills) underwent a standardized microbiological work-up including blood cultures, and cultures of sputum, throat swab, and urine. Diagnostic procedures for IA infection included mainly CT scans and sputum. BAL and other invasive procedures such as biopsy or needle aspiration were rarely performed. No assay for detecting fungal antigenemia was available during the study period. Patients did not receive antifungal prophylaxis covering Aspergillus spp. (Table 1).
Consecutive patients were recruited retrospectively by review of all charts of patients hospitalized for chemotherapy on the oncology ward during the study period. Patients were eligible for the study if they were hospitalized on the hemato-oncology ward at the University hospital Berne, received intensive chemotherapy for a hematological neoplasm, and were neutropenic (neutrophil count <500/μl and/or leukocyte count <1000/μl) between 1 January 1995 and 31 December 1999.
Cases of IA were defined as proven, probable or possible according to the recent recommendations.13 In addition, a fourth group of patients were defined as refractory fever cases, if they did not fulfill the criteria for IA, but presented with signs and symptoms of an infection not responding to treatment with broad spectrum antibiotics, no evidence for another infectious agents and clinical response to treatment with amphotericin B. Controls were patients without signs and symptoms of IA or refractory fever.
Data on clinical and sociodemographic characteristics were collected with a standardized questionnaire. The risk factors for IA analyzed were age, gender, the type of hematological neoplasm, duration of neutropenia, the number of chemotherapy induced neutropenic episodes during the preceding 6 months and the length of the time interval between the preceding (if applicable) and the current neutropenic episode. If a patient had more than one chemotherapy treatment during the study period, that is, for the first manifestation and/or relapse(s) of the malignancy, only one treatment was analyzed. For cases this was the treatment during which IA was diagnosed. For controls the first treatment was chosen. Risk factors for IA associated with chemotherapy and/or neutropenia were collected for the cycle immediately preceding the diagnosis of IA in cases and for the last chemotherapy cycle of the treatment in controls. The Charlson score served as a measure of comorbidity.14 This score is a method for classifying and weighting comorbid conditions (such as cardiopulmonal, renal or liver disease, etc), that may impact the risk of mortality associated with a primary diagnosis for admission to a hospital. In order to evaluate the ‘specificity’ of the risk factors for IA, the association of these factors with the risk of bacteremia was also analyzed.
Risk factors for IA were evaluated by univariate and multivariate analysis. Risk factors identified by univariate statistics were entered into a logistic regression model using StatView® version 5.0 (SAS Institute Inc., Cary, NC, USA). The final model contained the largest number of variables with P⩽0.05. Proportions were compared with the χ2 test or Fisher's exact test as appropriate. Differences between means were assessed by the Student's t-test. A cutoff of P⩽0.05, two tailed, was used for all statistical analyses.
A cohort of 189 patients was recruited into the study. The diagnosis of IA was proven in nine cases, probable in three cases, and possible in 33 cases. There were 11 cases with refractory fever and 133 controls. The lung was the site of infection in all proven, probable and possible cases. In 12 possible cases the diagnosis was based on a specific infiltrate in a CT scan (halo sign, air crescent sign or cavity within an area of consolidation),13 and in 21 possible cases symptoms of lower respiratory tract infection and a nonspecific pulmonary infiltrate were present. IA was diagnosed at autopsy in only one case. Death was attributed to IA in five of nine (55.5%) proven, one of three (33.3%) probable, seven of 33 (21.2%) possible IA cases and none of the patients with refractory fever. CT scans were performed more frequently in 1999. However, the proportion of possible IA cases did not increase in 1999 (eight of 13, 61%) as compared to earlier study years (25 of 32, 78.1%).
Risk factors for IA
Table 1 presents the sociodemographic and clinical characteristics of the study patients. Patients with IA differed significantly from controls for several variables. Leukemia or MDS were the underlying malignancies in 88% of IA cases as compared to 38% of controls (P<0.001). The prevalence of IA among the 101 study patients with acute leukemia or MDS was 8.9% for proven, 2.9% probable, 27.7% for possible IA cases and 9.9% had refractory fever. The duration of neutropenia was longer in IA cases than controls (mean 20.6 days vs 9.9 days, P<0.001). IA cases had a significantly lower number of chemotherapy induced neutropenic episodes during the previous 6 months than controls (mean 0.42 vs 1.03, P<0.001), that is, a high proportion (66.6%) of proven, probable and possible IA cases occurred during the first remission-induction chemotherapy cycle (Table 1). This percentage was significantly lower in refractory fever cases (27.3%, P=0.03) and controls (28.0%, P<0.001). Also, in study patients with at least one treatment cycle during the preceding 6 months, the time interval between the last day of the preceding and the first day of the index neutropenia tended to be shorter in IA cases than controls (mean 32.0 days vs 48.3 days, P=0.06). In the multivariate logistic regression model, acute leukemia and MDS, a longer duration of neutropenia and a lower number of preceding neutropenic episodes were independent risk factors for proven, probable and possible IA (Table 2). In the univariate analysis, a time interval of 14 days or less since the last neutropenic episode increased the risk of proven, probable or possible IA 4-fold (OR 4.3, CI 1.2–15.7, P=0.02). The OR did not change meaningfully in the multivariate model, but statistical significance was no longer reached due to small numbers (data not shown). When restricting the analysis to patients with acute leukemia or MDS, the associations between proven, probable and possible IA and longer duration of neutropenia and first chemotherapy cycle remained highly significant in the multivariate analysis. The mean duration of neutropenia in IA cases was 22.1 days vs 12.7 days in patients without IA (P<0.001). The proportion of IA cases, who were in their first cycle was 75.0% as compared to 31.3% for patients without IA (P<0.0001). The mean time interval since the last treatment cycle (for those patients with at least one preceding chemotherapy cycle) was shorter in IA cases (32.8 days) than control patients (21.0, P=0.19). Since only five possible and none of the proven and probable IA cases had lymphoma or myeloma, similar analyses could not be performed for this group of patients.
Of the 189 study patients, 51 (26.9%) had at least one bacteremia during the neutropenia. The proportions of blood cultures with a Gram positive (45.0%) or Gram negative bacteria (47.0%) were similar. Three (5.8%) patients had polymicrobial bacteremia with Gram positive and Gram negative germs, and one (1.5%) patient had candidemia. Patients with leukemia or MDS had a significantly higher risk of bacteremia than patients with lymphoma or myeloma (Table 3). The risk of a bacteremia was not related to the number of preceding chemotherapy cycles or the time interval since the last cycle (Table 3).
IA among neutropenic patients treated for hematologic malignancies represents a serious complication, the management of which must be based on a careful analysis of risk factors for this infection. The diagnosis of IA is difficult and remains tentative in a large proportion of suspected IA cases. Comparison of results from different studies has been hampered by an inconsistent use of case definitions for IA. Recently, consensus case definitions have been elaborated.13 In this study, we adopted these definitions. Since, in our institution, invasive diagnostic procedures were rarely performed and an assay for fungal antigenemia was not available during the study period, a fourth category was added named refractory fever cases. Possible IA and refractory fever cases were similar to proven and probable IA cases for the high proportion of acute leukemia or MDS and the long duration of neutropenia. They were therefore at increased risk of IA as discussed below. The high lethality for proven and possible IA (six of 12 cases) observed in this study was comparable to rates reported in recent series.2 The rates were lower for possible IA (21%) and refractory fever (0%) cases. This may in part be due to suspecting the diagnosis and beginning of treatment before the infection was fully developed in some of these patients; in addition, it is conceivable that not all patients in this group, especially the refractory fever group, had IA.
In accordance with earlier investigations patients with acute leukemia or MDS had the highest risk of IA (rate 12% for proven and probable IA).2 Denning recently reported a higher risk of IA in AML than ALL patients, but better survival in AML patients.15 In our study the rate of proven and probable IA was not higher in patients with AML (12%) than those with ALL (11%). Death occurred in four of 10 IA cases with AML and two of two cases with ALL, but these numbers were too small for drawing conclusions. Also, there was association of IA with the AML subtype (data not shown). No proven or probable IA was diagnosed among the 88 study patients with lymphoma or myeloma. Recently, a relatively high frequency of IA was found in a large multicenter European cohort of patients with myeloma.16 Our study cohort may have been too small to detect such IA cases. Also, our study confirmed the strong association between long duration of neutropenia and IA.2, 6
Bow et al10 found an increased risk of invasive fungal infection for a high-dose cytarabin regimen. However, their study included a high proportion of yeast infections. Our study patients received standardized treatment regimens according to their underlying malignancy. High-dose cytarabin was part of the treatment of relapse AML only and there was no obvious association between relapse AML and IA. Corticosteroids have been shown to increase the risk for IA.7, 8, 9 The quite uniform administration of corticosteroids to our study patients, mostly as part of an additive antiemetic prophylaxis, precluded to evaluate corticosteroids as an independent risk factor.
The risk of IA was significantly increased during the first or induction cycle of treatment. Wiley et al observed a higher incidence of invasive fungal infections during the induction cycle in pediatric patients. Most of these fungal infections were, however, caused by Candida species.17 In the study by Oren et al,18 IA occurred at equal frequency in the induction and consolidation cycle, but their study was not designed to analyze risk factors of IA. Pagano et al19 also observed a higher IA incidence (57%) in the first induction cycle, but they did not perform a multivariate analysis. In a recent review, Denning stated without reference that 75% of IA occur in the induction cycle vs 15% IA in the maintenance.2 The reason(s) for a higher risk of IA during the induction cycle are unclear. Wiley et al17 suggested that aplasia due to leukemia preceding the first treatment cycle may be an explanation. In our study, however, the association with the induction cycle did not depend on the duration of neutropenia. Ill-defined effects of the leukemic process on host defenses may act in concert with neutropenia to increase the risk of IA, which manifests at the earliest occasion, that is, the induction cycle when the leukemia is not yet treated. It may also be that exposure to aspergillus spores before hospital entry explains the high rate of IA during the first cycle.20 This is also supported by the lack of correlation between fungal strains found in and around the hospital environment and those causing infection in patients.21 The incubation time of IA is unknown. In neutropenic patients IA occurs predominantly after the 12th day of neutropenia6 despite upper respiratory tract colonization at entry.20 In our center, patients often remain hospitalized for subsequent treatment cycles and are therefore relatively protected from environmental exposure to spores. Lastly, there is evidence for acquired immunity against Aspergillus infection possibly mediated in part by macrophages.22 Whether such an immune response can be mounted during or following the first cycle that would then be protective during subsequent neutropenic episodes, is unclear.
A short time interval of less than 14 days between treatment cycles carried a higher risk for IA. This observation may be related to a qualitative disturbance of granulocyte function induced by the preceding chemotherapy cycle,11 or it may be indicative for poor response necessitating early institution of the next cycle of chemotherapy.
Older age has been associated with IA in bone marrow transplant (BMT) patients.23 In this study, which did not include BMT patients, proven and probable IA cases were not significantly older than controls, despite a wide age range in the study population. This may at least in part be related to the low prevalence of comorbidities, for which age may also serve as a surrogate. Indeed, in our institution, the age limits for curative treatment were 60 years for AML and ALL, and 65 years for MDS, multiple myeloma and lymphoma. Also, comorbidities were rare among our study population. Underlying comorbidities other than malignancies have to our knowledge not been studied systematically as risk factor. The Charlson score was designed as a predictor of mortality in cancer patients and covers a broad spectrum of diseases including diabetes mellitus, renal and liver insufficiency, cardiovascular and chronic lung disease.14 We found a very low average score in all patient groups and there was no correlation between comorbidities as measured by the Charlson score and the risk of IA. This result is important, since there is a general trend to offer curative treatment for hematological malignancy to patients of increasing age and with increasing comorbidities.
In conclusion, in this retrospective case control study, we found a significant association of IA with acute leukemia or MDS, duration of neutropenia, first or induction chemotherapy cycle and shorter time interval between treatment cycles. Older age and comorbidities were not predictive for IA. These results contribute to the further characterization of patient groups at high risk of IA and may help to target costly prophylactic measures against IA.
Bodey G, Bueltmann B, Duguid W, Gibbs D, Hanak H, Hotchi M et al. Fungal infections in cancer patients: an international autopsy survey. Eur J Clin Microbiol Dis 1992; 11: 99–109.
Denning DW . Invasive aspergillosis. Clin Infect Dis 1998; 26: 781–805.
Denning DW . Therapeutic outcome in invasive aspergillosis. Clin Infect Dis 1996; 23: 608–615.
von Eiff M, Roos N, Schulten R, Hesse M, Zuhlsdorf M, van de Loo J . Pulmonary aspergillosis: early diagnosis improves survival. Respiration 1995; 62: 341–347.
Prentice HG, Kibbler CC, Prentice AG . Towards a targeted, risk-based, antifungal strategy in neutropenic patients. Br J Haematol 2000; 110: 273–284.
Gerson SL, Talbot GH, Hurwitz S, Strom BL, Lusk EJ, Cassileth PA . Prolonged granulocytopenia: the major risk factor for invasive aspergillosis in patients with acute leukemia. Ann Int Med 1984; 100: 345–351.
Baddley JW, Stroud TP, Salzmann D, Pappas PG . Invasive mold infections in allogenic bone marrow transplant recipients. Clin Infect Dis 2001; 32: 1319–1324.
Roilides E, Uhlig K, Venzon D, Pizzo PA, Walsh TJ . Prevention of corticosteroid-induced suppression of human polymorphonuclear leukocyte-induced damage of Aspergillus fumigatus hyphae by granulocyte colony-stimulating factor and gamma interferon. Infect Immun 1993; 61: 4870–4877.
Ng TT, Robson GD, Denning DW . Hydrocortisone-enhanced growth of Aspergillus spp.: implications for pathogenesis. Microbiology 1994; 140: 2475–2479.
Bow EJ, Loewen R, Cheang MS, Schacter B . Invasive fungal disease in adults undergoing remission-induction therapy for acute myeloid leukemia: the pathogenetic role of the antileukemic regimen. Clin Infect Dis 1995; 21: 316–319.
Coiffier B, Frobert Y, Revol L . Polymorphonuclear function in acute myeloblastic leukemia. Biomedicine 1977; 27: 94–96.
Warnock DW, Hajjeh RA, Lasker BA . Epidemiology and prevention of invasive aspergillosis. Curr Infect Dis Rep 2001; 3: 507–516.
Ascioglu S, Rex JH, de Pauw B, Bennett JE, Bille J, Crokaert F et al. Defining opportunistic invasive fungal infections in immunocompromised patients with cancer and hematopoietic stem cell transplants: an international consensus. Clin Infect Dis 2002; 34: 7–14.
Charlson ME, Pompei P, Ales KL, MacKenzie CR . A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 1987; 40: 373–383.
Denning DW, Marinus A, Cohen J, Spence D, Herbrecht R, Pagano L et al. An EORTC multicentre prospective survey of invasive aspergillosis in haematological patients: diagnosis and therapeutic outcome. EORTC Invasive Fungal Infections Cooperative Group. J Infect 1998; 37: 173–180.
Lortholary O, Ascioglu S, Moreau P, Herbrecht R, Marinus A, Casassus P et al. Invasive aspergillosis as an opportunistic infection in nonallografted patients with multiple myeloma: a European Organization for Research and Treatment of Cancer/Invasive Fungal Infections Cooperative Group and the Intergroupe Francais du Myelome. Clin Infect Dis 2000; 30: 41–46.
Wiley JM, Smith N, Leventhal BG, Graham ML, Strauss LC, Hurwitz CA et al. Invasive fungal disease in pediatric acute leukemia patients with fever and neutropenia during induction chemotherapy: a multivariante analysis of risk factors. J Clin Oncol 1990; 8: 280–286.
Oren I, Haddad N, Finkelstein R, Rowe JM . Invasive pulmonary aspergillosis in neutropenic patients during hospital construction: before and after chemoprophylaxis and institution of HEPA filters. Am J Hematol 2001; 66: 257–262.
Pagano L, Girmenia C, Mele L, Ricci P, Tosti ME, Nosari A, et al., GIMEMA Infection Program, Gruppo Italiano Malattie Ematologiche dell’Adulto. Infections caused by filamentous fungi in patients with hematologic malignancies. A report of 391 cases by GIMEMA Infection Program. Haematologica 2001; 86: 862–870.
Martino P, Raccah R, Gentile G, Venditti M, Girmenia C, Mandelli F . Aspergillus colonization of the nose and pulmonary aspergillosis in neutropenic patients: a retrospective study. Haematologica 1989; 74: 263–265.
Leenders AC, van Belkum A, Behrendt M, Luijendijk A, Verbrugh HA . Density and molecular epidemiology of aspergillus in air and relationship to outbreaks of aspergillus infection. J Clin Microbiol 1999; 37: 1752–1757.
de Repentigny L, Petitbois S, Boushira M, Michaliszyn E, Senechal S, Gendron N et al. Acquired immunity in experimental murine aspergillosis is mediated by macrophages. Infect Immun 1993; 6: 3791–3802.
Marr KA, Carter RA, Crippa F, Wald A, Corey L . Epidemiology and outcome of mould infections in hematopoietic stem cell transplant recipients. Clin Infect Dis 2002; 34: 909–917.
About this article
Cite this article
Mühlemann, K., Wenger, C., Zenhäusern, R. et al. Risk factors for invasive aspergillosis in neutropenic patients with hematologic malignancies. Leukemia 19, 545–550 (2005). https://doi.org/10.1038/sj.leu.2403674
- invasive infection
Blood Reviews (2020)
Memórias do Instituto Oswaldo Cruz (2020)
Genetic Risk Surveillance for Invasive Aspergillosis in Hematology Patients: A Prospective Observational Study
Infectious Diseases and Therapy (2020)
Reconstituting NK Cells After Allogeneic Stem Cell Transplantation Show Impaired Response to the Fungal Pathogen Aspergillus fumigatus
Frontiers in Immunology (2020)
Can immune biomarkers predict infections in solid organ transplant recipients? A review of current evidence
Transplantation Reviews (2019)