Clinical significance of Aspergillus fungaemia in patients with haematological malignancies and invasive aspergillosis

Authors


Corrado Girmenia, MD, Dipartimento di Biotecnologie Cellulari ed Ematologia, University ‘La Sapienza’, Via Benevento 6, 00161 Rome, Italy. E-mail: girmenia@bce.med. uniroma1.it

Abstract

The clinical significance of Aspergillus fungaemia in the setting of a deep-seated aspergillosis has not been clearly established. Among 107 microbiologically documented Aspergillus infections in patients with haematological diseases observed over a 17-year period, blood cultures grew Aspergillus species from 10 cases. Aspergillus fungaemia was documented in 9 out of 89 (10·1%) patients with pulmonary aspergillosis at a median of 5 d from the onset of clinical signs of infection, and in one patient with central venous catheter focal infection. Five (50%) patients died as a result of fungal infection a median of 12 d (range 4–48) from the documentation of Aspergillus fungaemia. A comparison between cases of invasive aspergillosis with or without fungaemia showed that fungaemic patients were similar to those without positive blood cultures regarding clinical presentation, risk factors, clinical course and outcome. The diagnostic role of Aspergillus fungaemia in the setting of a deep-seated infection is limited because blood cultures become positive when a microbiological or clinical diagnosis of aspergillosis has already been performed. Aspergillus fungaemia does not necessarily seem to be correlated with a disseminated infection or a poorer prognosis.

Blood cultures represent a limited diagnostic tool for invasive aspergillosis because Aspergillus fungaemia is rarely encountered, even in the setting of disseminated disease (Duthie & Denning, 1995; Denning, 1998). Furthermore, determining the importance of Aspergillus fungaemia in the immunocompromized patient is difficult because media contamination, yielding false-positive results, of cultures can occur.

Criteria for differentiating true Aspergillus fungaemia from contamination have been proposed by Duthie & Denning (1995). By reviewing the cases of Aspergillus fungaemia reported in the literature, they observed that true Aspergillus fungaemia occurred more frequently after cardiac surgery or in patients with neutropenia. In a recently reported 10-year single-centre retrospective study on the significance of fungaemia as a result of Aspergillus species in patients with cancer, cases of true Aspergillus fungaemia were seen exclusively in patients with haematological malignancies in late phases of invasive fungal infection (Kontoyiannis et al, 2000). However, the clinical significance and the prognostic value of Aspergillus fungaemia in the setting of an invasive disease have not been clearly established.

In this paper, we report a 17-year single-centre retrospective study of microbiologically documented Aspergillus infections in patients with haematological diseases, designed to characterize the clinical aspects and prognosis of Aspergillus fungaemia in this category of immunocompromized subjects. We also reviewed all cases of Aspergillus fungaemia in patients with haematological diseases reported in the literature.

Patients and methods

We retrospectively reviewed the charts of all patients with haematological diseases who had a microbiologically documented Aspergillus infection between October 1982 and December 1999 at the Institute of Haematology, a 60-bed teaching hospital of the University ‘La Sapienza', Rome, Italy. Patients were admitted to single- or double-bedded rooms and received reverse isolation precautions. Norfloxacin or ciprofloxacin, in combination with oral amphotericin B, were given to all patients as prophylaxis. Until 1990, patients also received a nasal spray of amphotericin B as prophylaxis for Aspergillus species colonization of the nose and sinuses. In the event of fever and neutropenia (< 0·5 × 109 polymorphonuclear leucocytes/l), blood cultures were collected and the patients were started on empirical antibiotics, usually a β-lactam plus an aminoglycoside. Blood cultures were performed using trypticase soy broth bottles (BCG System, Roche, and Sygnal System, Hants, UK), incubated at 37°C and examined daily for at least 2 weeks. Broth bottles suspected to be positive at visual examination were subcultured onto blood agar. When there was the suspicion of a fungal growth, subculture onto Sabouraud's dextrose agar was also performed. Semiquantitative cultures of the tips of removed central venous catheters (CVCs) were performed according to reported procedures (Maki et al, 1977). Surveillance cultures of urine, stool and sputum specimens, as well as of nasal, oropharyngeal, rectal and vaginal swabs, were performed weekly, plated onto Sabouraud's dextrose agar and incubated at 37°C for 10 d. Identification of Aspergillus species was performed using standard morphological methods. Amphotericin B (1 −1·5 mg/kg/d) was administered to patients with unexplained fever who were unresponsive to empirical antibacterial therapy and to patients with microbiological and/or clinical diagnosis of aspergillosis.

An episode of Aspergillus fungaemia was defined according to previously reported criteria (Duthie & Denning, 1995). Definite Aspergillus fungaemia was defined as the growth of Aspergillus species from blood cultures associated with either histological evidence of hyphae in tissue from other sites or positive cultures for the same Aspergillus species in the setting of a clinical condition compatible with invasive aspergillosis (i.e. pulmonary infiltrate and Aspergillus isolation from sputum culture). Catheter-related definite Aspergillus fungaemia was considered if the clinical setting was not compatible with invasive Aspergillus infection, the catheter-tip culture yielded Aspergillus species, and blood cultures became negative after CVC removal. Consequently, the definition of catheter-related Aspergillus fungaemia required removal of the CVC. Probable Aspergillus fungaemia was defined as the growth of Aspergillus species from blood cultures in a patient with clinical signs of aspergillosis, but without mycological or histopathological evidence of invasive aspergillosis (i.e. pulmonary infiltrate requiring amphotericin B therapy). Aspergillus pseudofungaemia was defined as the isolation of Aspergillus species from blood cultures in an atypical clinical setting or with a clinical course not compatible with invasive or CVC-related Aspergillus infection. Seven of our cases of Aspergillus fungaemia have been previously reported (Martino et al, 1993; Girmenia et al, 1995).

We performed a MEDLINE search of the English-language literature regarding Aspergillus fungaemia in patients with haematological diseases. References in the articles were reviewed to identify additional cases.

In order to identify factors associated with Aspergillus fungaemia and to evaluate its prognostic importance, patients with Aspergillus fungaemia were compared with those with invasive aspergillosis without fungaemia regarding age, sex, underlying disease and status, bone marrow transplant, presence of a CVC, previous use of steroids, presence and duration of neutropenia, clinical presentation of invasive aspergillosis, species of Aspergillus, antifungal prophylaxis, treatment and outcome. For the comparison between dichotomous variables, we used Fisher or chi-square tests as appropriate. For continuous variables we used the Wilcoxon test.

Results

One hundred and seven cases of Aspergillus infection in patients with haematological diseases were microbiologically documented at our Institution over the study period. The characteristics of the patients and the infections are described in Tables I and II respectively. Pulmonary involvement was documented in 89 patients (83%). It was the primary site of infection in 83 cases (77·6%) and the single documented localization of infection in 67 patients (62·6%). Twenty-one patients had sinusitis (19·6%); it was the primary site of infection in 19 cases (17·8%) and the single documented localization of infection in 16 patients (15%).

Table I.   Characteristics of 107 patients with microbiologically documented Aspergillus infections.
VariableNumber
Number of patients107
Median age in years (range)39 (4–75)
Male:female ratio79:28
Underlying disease
 Acute myeloid leukaemia56
 Acute lymphoid leukaemia19
 Chronic myeloid leukaemia, blast crisis14
 Non-Hodgkin's lymphoma7
 Aplastic anaemia4
 Other haematological malignancies7
Phase of the underlying disease
 Onset50
 Complete remission34
 Relapse22
 Resistant1
Treatment for the underlying disease
 Chemotherapy77
 Allogeneic bone marrow transplantation17
 Autologous bone marrow transplantation5
 Other8
Table II.   Characteristics of microbiologically documented invasive Aspergillus infections in 107 patients.
VariableNumber (%)
  • *

    The sum exceeds 107 because many patients had more than one site involved.

  • One organ involved.

  • Two or more organs involved.

  • §

    Only microbiological data associated with local clinical evidence of invasive Aspergillus infection were considered.

  • In seven patients, microbiological diagnosis of aspergillosis was achieved before death.

Site involved*
 Lungs89 (83·2)
 Sinuses21 (19·6)
 Brain7 (6·5)
 Skin6 (5·6)
 Gastrointestinal tract3 (2·8)
 Bone2 (1·9)
 Central venous catheter4 (3·7)
Primary site of infection
 Lungs83 (77·6)
 Sinuses19 (17·8)
 Skin2 (1·9)
 Central venous catheter exit site2 (1·9)
 Central venous catheter tip1 (0·9)
Localized infections 85 (79)
Disseminated infections22 (21)
Sites of positive cultures for Aspergillus species*§
 Sputum66(61·7)
 Bronchoalveolar lavage5 (4·7)
 Lung biopsy1 (0·9)
 Bronchial biopsy1 (0·9)
 Blood10 (9·3)
 Nasal swab18 (16·8)
 Sinus biopsy2 (1·9)
 Skin biopsy8 (7·7)
 Bone marrow1 (0·9)
 Secretion from bone fistula2 (1·9)
 Central venous catheter tip2 (1·9)
 Central venous catheter exit site2 (1·9)
 Specimens obtained from autopsy13 (12·1)

In 10 of the 107 (9·3%) cases of microbiologically documented invasive aspergillosis, blood cultures grew Aspergillus species. The characteristics of the cases of Aspergillus fungaemia are detailed in Table III. There were nine cases of definite and one case of probable Aspergillus fungaemia. In addition, three cases of pseudofungaemia were identified during the study period.

Table III.   Characteristics of 10 cases of true Aspergillus fungaemia in patients with haematological malignancies.
Case
number
Age (years)
/Sex

Classification
Underlying
disease
Treatment for the
underlying disease
Organ(s)
involved
Positive culture specimens
from other sites

Species
Number of +ve blood
cultures/d of fungaemia

Treatment

Outcome
  1. *Cases previously reported (Martino et al, 1993).

  2. †Case previously reported (Girmenia et al, 1995).

  3. AML, acute myeloid leukaemia; ALL, acute lymphoid leukaemia; AmB, amphotericin B; BAL, bronchoalveolar lavage; BMT, bone marrow transplantation; CML, chronic myeloid leukaemia; CVC, central venous catheter.

1*22/MaleDefiniteALLChemotherapyLung, skinSkinA. fumigatus4/2AmBSurvived
2*33/MaleDefiniteAMLChemotherapyLungSputumA. fumigatus2/2AmBSurvived
3*50/FemaleDefiniteAMLChemotherapyLungSputumA. fumigatus2/2AmBSurvived
4*43/MaleDefiniteAMLChemotherapyLungSputum, lung (autopsy)A. fumigatus2/1AmBDied (32 d from
fungaemia)
5*28/FemaleDefiniteAMLChemotherapyLungLung (autopsy)A. fumigatus3/3AmBDied (4 d from
fungaemia)
6*35/MaleDefiniteCMLAllogeneic BMTLungLung (autopsy)A. fumigatus4/5AmBDied (6 d from
fungaemia)
726/FemaleDefiniteALL, HIV+ChemotherapyCVC, skin (CVC
exit site), lung
Skin (CVC exit site),
sputum, CVC tip
A. fumigatus6/4Itraconazole,
AmB
Died (48 d
from fungaemia)
830/FemaleDefiniteAMLAllogeneic BMTLungBALA. fumigatus1/1AmBDied (12 d from
fungaemia)
938/FemaleProbableAMLChemotherapyLungNoneA. flavus1/1AmBSurvived
104/MaleDefiniteALLChemotherapyCVCCVC tipA. flavus3/6NoSurvived

Aspergillus species was isolated in blood cultures in nine patients with pulmonary infection (10·1% of 89 cases of pulmonary aspergillosis) and in one patient with a CVC Aspergillus infection without any evidence of invasive disease. In the nine patients with pulmonary infection, positive blood cultures were documented at a median of 5 d (range 1–12) from the onset of clinical signs of invasive aspergillosis, and four patients were already under amphotericin B therapy. The only clinical manifestation of infection in the patient with CVC-related infection was mild fever. Three blood cultures taken from the CVC over a 6 d period were positive for Aspergillus flavus, whereas all four blood cultures obtained from peripheral vein puncture were negative. The CVC was removed and culture of the catheter tip yielded A. flavus. Fungaemia was cleared after CVC removal, no antifungal treatment was administered (the patient was not neutropenic) and no other sign of Aspergillus infection was observed. The median number of positive blood cultures was 2·5 (range 1–6) and the median duration of fungaemia was 2 d (range 1–6). Five (50%) patients died as a result of fungal infection a median of 12 d (range 4–48) from the documentation of Aspergillus fungaemia.

The comparison between patients with invasive aspergillosis who had and who did not have fungaemia is shown in Table IV. Aspergillus fungaemia was significantly associated with a lower duration of steroids use before diagnosis (P = 0·05). In addition, positive cultures from a catheter tip or from CVC exit site were more frequent in patients with Aspergillus fungaemia (2 out of 10 versus 2 out of 97, P = 0·04). Aspergillus fungaemia was not associated with disseminated infection (2 out of 10 cases with fungaemia versus 20 out of 97 cases without fungaemia, P = 1·0) and the death rates of both groups were similar (5 out of 10 versus 59 out of 97, P = 0·52).

Table IV.   Comparison between patients with invasive aspergillosis who had and who did not have Aspergillus fungaemia.

Variable
Fungaemia
n = 10
No fungaemia
n = 97

P-value
Age (years), median (range)31·5 (4–50)40 (11–75)0·09
Sex, male:female5:574:230·12
Underlying disease
 Acute non-lymphocytic leukaemia6500·74
 Acute lymphoid leukaemia3160·38
 Chronic myeloid leukaemia1131·0
 Aplastic anaemia041·0
 Non-Hodgkin's lymphoma071·0
 Other haematological malignancies071·0
Status of the underlying disease
 Initial treatment3470·33
 Complete remission4300·72
 Relapse/resistant3200·44
Bone marrow transplant2201·0
 Allogeneic2150·66
 Autologous051·0
Central venous catheter7681·0
Duration of steroid treatment before diagnosis, median (range)7·5 (0–60)25 (0–400)0·05
Number of neutrophils at diagnosis (× 109/l), median (range)0·8 (0·02–9·99)0·11 (0–9·99)0·34
Duration of profound (< 0·1 × 109/l) neutropenia before diagnosis, median (range)8 (0–15)8 (0–40)0·37
Duration of neutropenia (< 0·5 × 109/l) before diagnosis, median (range)11 (0–17)12 (0–200)0·21
Pneumonia9801·0
 Single lesion6420·5
 Multiple lesions338 
Sinusitis0210·20
Other sites of infection3200·44
 Central venous catheter220·04
 Liver011·0
 Bowel041·0
 Bone021·0
 Skin160·51
 Central nervous system071·0
Disseminated infection2201·0
Aspergillus species
 A. fumigatus8/1048/651·0
 A. flavus2/1010/650·66
 A. terreus06/651·0
 A. niger01/651·0
Antifungal prophylaxis
 Oral amphotericin B3380·74
 Itraconazole131·0
 Amphotericin B nasal spray061·0
 No prophylaxis6500·74
Treatment
 Amphotericin B9811·0
 Itraconazole151·0
 No treatment1111·0
 Death5590·52

Discussion

Aspergillus fungaemia seems to be a rare infection, even in patients with haematological malignancies, although neutropenia represents one of the most frequent underlying conditions for Aspergillus fungaemia (Duthie & Denning, 1995). To our knowledge, including the present series, 32 cases of true Aspergillus fungaemia in patients with haematological diseases have been reported to date (Winston et al, 1979; Hoy et al, 1986; Denning & Williams, 1987; Robertson & Larson, 1988; Schonheyder et al, 1992; Martino et al, 1993; Girmenia et al, 1995; Berner et al, 1996; Atra et al, 1998; Schett et al, 1998; Kontoyiannis et al, 2000). In addition, only three cases of Aspergillus fungaemia have been reported in solid tumours (Viollier et al, 1986; Gucalp et al, 1991; Marec-Berard et al, 1996). Acute leukaemia was the most frequent underlying condition (75% of cases). Aspergillus fumigatus was the most frequently isolated agent (16 out of 29 cases with available species identification, 55%), followed by Aspergillus terreus (eight cases, 27·6%), Aspergillus flavus (three cases, 10·3%) and Aspergillus niger (two cases, 6·9%). All but one case of fungaemia caused by A. terreus were clustered in a single institution (Kontoyiannis et al, 2000). The lungs were the primary site of infection in 78% of cases. Central venous catheter, with (two cases) or without (three cases) documented secondary deep-seated infection, was considered the presumed portal of entry of Aspergillus fungaemia 15·6% of cases. Death occurred in 55% of cases.

Twelve cases of definite or probable Aspergillus fungaemia, all in patients with haematological malignancies, were reported in a recent 10-year retrospective study performed at the M.D. Anderson Cancer Center (Kontoyiannis et al, 2000). It represents the largest single centre series reported to date. However, in that study, seven (58%) cases were documented by positive blood cultures taken at autopsy. Therefore, our study represents the largest series of true Aspergillus fungaemia documented before death.

The frequency of true Aspergillus fungaemia in patients with haematological malignancies seems to be low. In a recent prospective surveillance study of fungaemia in cancer patients conducted by the Invasive Fungal Infection Cooperative Group (IFICG) of the European Organization for Research and Treatment of Cancer (EORTC) (Viscoli et al, 1999), five case of Aspergillus fungaemia (1·9% of 270 fungaemia episodes) were observed among 30 centres over a 2-year period (unpublished data). In the present series, Aspergillus fungaemia accounted for 4% of all fungaemias (10 out of 226 cases) and for 9·3% of all microbiologically documented cases of invasive aspergillosis diagnosed in the 17-year study period. From these limited data, it seems that the incidence of Aspergillus fungaemia is around 2–4% of all fungaemias. Data from prospective multicentre studies conducted in institutions with high documentation of invasive aspergillosis are needed in order to estimate the rate of true Aspergillus fungaemia.

Ten percent of our patients with pulmonary aspergillosis were fungaemic, whereas no patient with sinusitis had the fungus isolated from blood. Aspergillus fungaemia had a minimal impact on the early diagnosis of aspergillosis, as all nine patients with pneumonia had a clinical or microbiological diagnosis of aspergillosis by the time the blood cultures become positive. Similarly, Aspergillus fungaemia was not a late occurrence in the course of the invasive infection as it was documented at a median of 5 d from the onset of clinical signs and, in the five patients who did not survive, at a median of 12 d before death. These data are in contrast to those reported by Kontoyiannis et al (2000), who observed that positive blood cultures occurred late in the course of infection (1 d before death or at autopsy).

The clinical significance of Aspergillus fungaemia in immunocompromized patients is uncertain as the prognostic value of positive blood cultures for Aspergillus species in the setting of invasive disease has not been clearly established. We analysed the clinical characteristics and mortality associated with Aspergillus fungaemia by comparing patients with invasive aspergillosis who had or did not have fungaemia. A significant association of Aspergillus fungaemia with a lower duration of previous steroid treatment was observed. We do not have a clear interpretation for this. Influence of steroids in favouring thrombotic fungal invasion with distal infarction and consequent reduced blood flow through the affected areas could be hypothesized. CVC involvement (exit site infection or catheter tip positive culture) was significantly associated with Aspergillus isolation from blood. With these exceptions, we found that patients with fungaemia were similar to those without positive blood cultures regarding clinical presentation, risk factors, clinical course and outcome. Importantly, Aspergillus fungaemia was not associated with disseminated infection or increased risk of death.

In conclusion, true Aspergillus fungaemia seems to be a rare event even in high-risk patients with haematological malignancies. It is usually documented in patients with pulmonary aspergillosis and, in a minority of cases, in patients with CVC Aspergillus infection. Its diagnostic role in the setting of a deep-seated infection is limited because blood cultures usually become positive when a microbiological or clinical diagnosis of aspergillosis has already been performed. Isolation of Aspergillus from blood in a patient with invasive aspergillosis does not necessarily seem to represent a sign of disseminated infection or be correlated with a poorer prognosis.

Ancillary