Multiple-species candidemia in patients with cancer

Authors

  • Maha R. Boktour M.D.,

    1. Department of Infectious Diseases, Infection Control and Employee Health, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Dimitrios P. Kontoyiannis M.D., Sc.D.,

    Corresponding author
    1. Department of Infectious Diseases, Infection Control and Employee Health, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    • Department of Infectious Diseases, Infection Control and Employee Health, Unit 402, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Boulevard, Houston, TX 77030
    Search for more papers by this author
    • Fax: (713) 745-6839

  • Hend A. Hanna M.D., M.P.H.,

    1. Department of Infectious Diseases, Infection Control and Employee Health, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Ray Y. Hachem M.D.,

    1. Department of Infectious Diseases, Infection Control and Employee Health, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Essam Girgawy M.D.,

    1. Department of Infectious Diseases, Infection Control and Employee Health, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Gerald P. Bodey M.D.,

    1. Department of Infectious Diseases, Infection Control and Employee Health, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author
  • Issam I. Raad M.D.

    1. Department of Infectious Diseases, Infection Control and Employee Health, The University of Texas M. D. Anderson Cancer Center, Houston, Texas
    Search for more papers by this author

  • Presented in part as an abstract at the 41st Annual Meeting of the Infectious Diseases Society of America, San Diego, CA, October 9–12, 2003.

Abstract

BACKGROUND

Candidemia is a common cause of bloodstream infections in patients with cancer, with the majority of these infections being caused by a single Candida species. Studies of multiple-species candidemia (MSC) have rarely been reported.

METHODS

The authors identified 33 patients with cancer who had candidemia (diagnosed between 1993 and 2000) caused by more than 1 Candida species. This group of 33 patients was compared with a control group of 66 patients with cancer who had C. albicans candidemia that arose soon before or soon after each case of MSC that was investigated in the current study.

RESULTS

Patients with MSC, compared with control patients, were more likely to have leukemia (33% vs. 8%; P = 0.001), to have had prolonged neutropenia before the onset of their infection (mean ± standard deviation, 10 ± 17 days vs. 3 ± 6 days; P = 0.02), and to have received chemotherapy within 1 month before their infection (42% vs. 18%; P = 0.01). Patients with MSC also had higher Acute Physiology and Chronic Health Evaluation II scores at the onset of infection (score ≥ 16, 45% vs. 26%; P = 0.05) and were more likely to have received previous antifungal prophylaxis compared with patients who had candidemia caused by C. albicans (33% vs. 11%; P = 0.006). The response of C. albicans candidemia to single-agent antifungal therapy was significantly better than that of MSC (69% vs. 35% P = 0.004).

CONCLUSIONS

In patients with cancer, MSC was more likely to occur as breakthrough candidemia, predominantly in those with leukemia and prolonged neutropenia, and was associated with suboptimal responses to single-agent antifungal therapy. Cancer 2004. © 2004 American Cancer Society.

The incidence of invasive fungal infections has increased over the past two decades, especially in immunosuppressed and critically ill patients, and these infections are associated with high morbidity and mortality rates.1–4 Candidemia is the fourth most common cause of bloodstream infections in the United States.2, 3 For many years, Candida albicans accounted for the majority of candidemia cases.4, 5 However, non–C. albicans species have recently emerged as major pathogens in patients receiving azoles either as prophylactic or empiric therapy, especially those being treated for acute leukemia and those who have received bone marrow transplants.1, 3, 6–10 Because most candidemias are caused by a single Candida species,5, 8–19 attention has been focused on the clinical features of these cases. In contrast, multiple-species candidemias (MSC) are infrequent, accounting for 2–10% of cases in large series.8–19 Very few studies have focused on MSC.20–22 Therefore, we conducted the current retrospective study, which covered a 6-year period, to identify risk factors and other relevant prognostic data associated with MSC in patients with cancer, and we compared our findings with those associated with C. albicans candidemia in the same population.

MATERIALS AND METHODS

By reviewing the microbiology records of the University of Texas M. D. Anderson Cancer Center (Houston, TX), we identified 33 patients with cancer who developed MSC between January 1993 and January 2000. Routine methods for the isolation and speciation of Candida species were followed.23 These 33 patients were compared with 66 patients with cancer who had C. albicans candidemia, with each C. albicans infection having occurred within 1 month before or after each case of MSC (2 cases of C. albicans infection for every case of MSC). A patient with MSC was defined as having ≥ 2 Candida species in 1 or more blood culture specimens obtained either from the initial blood culture or within 72 hours from the appearance of signs and symptoms of systemic infection. Patients' medical reports were reviewed for the following demographic characteristics and potential risk factors: age, gender, underlying disease, and Acute Physiology and Chronic Health Evaluation (APACHE) II score24 (≥ 16 or < 16) on the day of the first candidemia-positive blood culture (index culture). In addition, information regarding the duration of hospital and intensive care unit (ICU) stays, previous adrenal corticosteroid use, chemotherapy use, presence or absence of a central venous catheter, occurrence of total parental hyperalimentation, prophylactic antifungal use, incidence of neutropenia, response to antifungal therapy, and total mortality was collected. These clinical characteristics were noted within 30 days of the index blood culture and also during the course of the infection. Response to antifungals and outcome were evaluated at 30 days after the first blood culture was obtained.

Definitions

Breakthrough candidemia was defined as candidemia occurring during the administration of systemic antifungal agents as prophylactic or empiric therapy and within 7 days before the onset of infection.7 Neutropenia was defined as an absolute neutrophil count < 500/mm3 at the onset of infection. Response to antifungal treatment was defined as the resolution of all clinical manifestations of candidemia, with follow-up blood cultures that were negative for Candida species, whereas antifungal treatment failure was defined as the persistence or progression of the clinical signs and symptoms of the fungal infection and/or the continued finding in blood cultures of the same Candida species that were originally present.

Statistical Methods

For categoric data, we used the Fisher exact test to compare the MSC and C. albicans cohorts and to evaluate the response of MSC to antifungals. For numeric data, we used the Student t test and the Wilcoxon rank-sum test. Also, we used logistic regression analysis with stepwise backward elimination to identify multiple discriminators between MSC and C. albicans and predictors of MSC response. We included only relevant parameters that had a value of P ≤ 0.05 in the univariate analysis. P ≤ 0.05 was considered indicative of statistical significance, and no adjustments were made for multiple comparisons. All analyses were performed using the statistical computing package SPSS (Version 11.0; SPSS, Inc., Chicago, IL).

RESULTS

We identified 33 patients with MSC during the study period, which represented (6%) of all patients with candidemia. Eleven of the 33 patients with MSC (33%) had an underlying hematologic malignancy, predominantly acute leukemia (Table 1). In contrast, the majority of the 66 patients with C. albicans candidemia (77%) had an underlying solid tumor (P = 0.009). A greater proportion of patients with MSC had baseline APACHE II scores ≥ 16 (45% vs. 26%; P = 0.05). Neutropenia was present at the onset of candidemia in 58% of patients with MSC compared with 27% of patients with with C. albicans infection (P = 0.003). Also, the duration of neutropenia before the onset of candidemia was longer in the MSC group (mean ± standard deviation, 10 ± 17 days vs. 3 ± 6 days; P = 0.02). In addition, patients with MSC were more likely to have remained neutropenic during their infection (30% vs. 8%; P = 0.005) and to have received growth factor support (46% vs. 24%; P = 0.007; Table1).

Table 1. Patient Demographics and Clinical Characteristics
VariableNo. of patients (%)P valuea
MSC (n = 33)Candida albicans candidemia (n = 66)
  • SD: standard deviation; NS: not statistically significant; BMT: bone marrow transplantation; MDS: myelodysplastic syndrome; ICU: intensive care unit; MSC: multiple-species candidemia; APACHE: Acute Physiology and Chronic Health Evaluation.

  • a

    Univariate analysis.

  • b

    Calculated as the duration between the first and last candidemia-positive blood cultures.

  • c

    Within 30 days before the first candidemia-positive blood culture or at study entry.

  • d

    At study entry (i.e., on the day of the first candidemia-positive blood culture).

Mean age ± SD (yrs)53 ± 2051 ± 18NS
Males/females17/1630/36NS
Acute leukemia or MDS11 (33) 5 (8)0.001
Solid tumor17 (52)51 (77)0.009
Lymphoma/melanoma 5 (15)10 (15)NS
BMT within past year 4 (12) 2 (3)NS
ICU stay during infection16 (48)29 (44)NS
Mean duration of ICU stay ± SD (days)12 ± 1412 ± 12NS
Duration of candidemia ≥ 5 daysb16 (48)13 (20)0.003
Mean duration of candidemia episode ± SD (days)5.0 ± 5.02.7 ± 4.00.005
Neutropenia (< 500/mm3)c13 (39)16 (24)NS
Neutropenia (< 500/mm3)d19 (58)18 (27)0.003
Persistent neutropenia (vs. recovery or no neutropenia)10 (30) 5 (8)0.005
Mean duration of neutropenia ± SD (days)d10 ± 173 ± 60.02
Growth factor usec15 (46)16 (24)0.03
Chemotherapy usec23 (70)28 (42)0.01
Corticosteroid usec14 (42)19 (29)NS
APACHE II score ≥ 16d15 (45)17 (26)0.05
Antifungal therapy during previous month11 (33) 7 (11)0.006
Fluconazole prophylaxis 9 (27) 7 (11)0.03
Breakthrough candidemia 6 (18) 3 (5)0.04
Catheter-related candidemia10 (30)25 (38)NS

More patients with MSC received systemic antifungals as either empiric or prophylactic treatment during the month before their first positive blood culture (33% vs. 11%; P = 0.006). Fluconazole was the most common antifungal used during the month preceding the onset of MSC (27% vs. 11%; P = 0.03). Breakthrough candidemia was more common in patients with MSC (18% vs. 5%; P = 0.04). Forty-eight percent of patients with MSC had positive blood cultures for ≥ 5 days, compared with only 20% of patients with C. albicans candidemia (P = 0.003). The mean duration of candidemia was longer in the MSC cohort as well (5 days vs. 4 days; P = 0.005). There was no difference in the frequency of catheter-related candidemia between the two groups (Table 1). In multivariate analysis, the presence of leukemia (odds ratio [OR], 6.25; 95% confidence interval [CI], 1.9–20.40) and an APACHE II score ≥ 16 (OR, 2.47; 95% CI, 1.00–6.30) were independent factors favoring the occurrence of MSC.

C. albicans was the most common species cultured from the blood specimens of patients with MSC (21 of 33 [64%]), followed by C. glabrata (55%), and C. tropicalis (42%). The most common combinations of Candida species were C. albicans and C. glabrata (9 of 33 [27%]), C. albicans and C. tropicalis (5 of 33 [15%]), and C. albicans and C. parapsilosis (5 of 33 [15%]; Table 2). Three patients had three distinct Candida species isolated from blood cultures.

Table 2. Number of Patients with MSC by Candida Species Combination
Candida speciesNo. of patientsPersistent species at autopsy (n = 6)
Total (n = 33)Experienced treatment failure (n = 18)
  1. MSC: multiple-species candidemia.

C. albicans + C. glabrata96C. albicans, C. albicans, and C. glabrata
C. albicans + C. tropicalis52
C. albicans + C. parapsilosis52
C. albicans + C. guillermondi11
C. tropicalis + C. glabrata31
C. tropicalis + C. krusei11C. krusei
C. tropicalis + C. parapsilosis2
C. glabrata + C. krusei33C. krusei
C. glabrata + C. parapsilosis1
C. albicans + C. tropicalis + C. glabrata11C. tropicalis and C. glabrata
C. tropicalis + C. krusei + C. lusitaniae1
C. tropicalis + C. glabrata + C. krusei11C. tropicalis and C. krusei

Ninety-one percent of patients with MSC received antifungal therapy (amphotericin B [AMB] deoxycholate [n = 4], lipid formulations of AMB [n = 10], or fluconazole [n = 12]) for their infection (Table 3). The overall clinical and mycologic response was only 47% in patients with MSC, compared with 74% in patients with C. albicans (14 of 30 [47%] vs. 40 of 54 [74%]; P = 0.01). The response rate in patients with MSC who received a single antifungal agent as initial therapy was only 35% (9 of 26). In contrast, in the 4 patients with MSC who received combination antifungal therapy consisting of an AMB formulation plus fluconazole at the onset of infection, the response was 75% (3 of 4). The 30-day overall mortality rate for patients with MSC was 45%, compared with 33% for patients with C. albicans candidemia, a difference that was not statistically significant.

Table 3. Therapy and Outcome Data for Patients with MSC and Patients with Candida albicans Candidemia
VariableNo. of patients (%)P valuea
MSC (n = 33)C. albicans candidemia (n = 66)
  • AMB: amphotericin B; NS: not statistically significant; MSC: multiple-species candidemia.

  • a

    Univariate analysis.

Antifungal therapy use30 (91)54 (82)NS
 Fluconazole12370.02
 AMB deoxycholate47NS
 Lipid formulation of AMB1070.05
 Combination (polyene plus fluconazole) 4 (12) 3 (5)NS
Had response to antifungal therapy14 (47)40 (74)0.01
 Had response to primary single-agent therapy 9 (35)35 (69)0.004
 Had response to combination therapy (polyene plus fluconazole) 3 (75) 2 (67)NS
30-day mortality15 (45)22 (33)NS

The factors influencing outcome for patients with MSC are shown in Table 4. The treatment failure rate in the MSC group was 91% for patients with underlying leukemia or myelodysplastic syndrome (P = 0.003), 77% for patients who had neutropenia 1 month before their first positive blood culture (10 of 13; P = 0.04), and 74% for patients who were neutropenic at the onset of their infection (14 of 19; P = 0.01). Similarly, the treatment failure rate was higher in patients with MSC who had persistent neutropenia (90% [9 of 10]) than in those who had recovered from neutropenia (P = 0.008). Poor outcomes were observed in all three patients with C. glabrata/C. krusei MSC and in two of the three patients who had at least three Candida species isolated from their blood cultures. C. krusei was the most common Candida species found in the six patients who underwent autopsy (3 of 6). Patients with MSC who required ICU admission during their infection had a higher rate of treatment failure (75% [12 of 16]) than those who did not (25% [4 of 16]; P = 0.02). In contrast, MSC responded to antifungal treatment in the majority of patients who had an underlying solid tumor (80%). Underlying leukemia was the only independent factor predicting poor outcome for patients with MSC on multivariate analysis (OR, 17.5; 95% CI, 19.0–166.7; P = 0.012).

Table 4. Univariate Analysis of Predictors of Poor Outcome in Patients with MSC
FactorNo. of patients (%)P value
Response to therapy (n = 15)Treatment failure (n = 18)
  • MDS: myelodysplastic syndrome; MSC: multiple-species candidemia; ICU: intensive care unit.

  • a

    Within 30 days before the first candidemia-positive blood culture.

  • b

    At study entry (i.e., on the day of first candidemia-positive blood culture).

  • c

    Less than 500/mm3 throughout the study period.

Underlying disease   
 Leukemia or MDS1 (7)10 (56)0.003
 Solid tumor12 (80)5 (28)0.003
Neutropenia (< 500/mm3)a3 (20)10 (56)0.04
Growth factor usea3 (20)12 (67)0.007
Neutropenia (< 500/mm3)b5 (33)14 (79)0.01
Persistent neutropeniac1 (7)9 (50)0.008
ICU stay during infection4 (27)12 (67)0.02

DISCUSSION

MSC occurs in a minority of patients with candidemia.5, 8–19 This infection has rarely been investigated and has received little attention. Herein, we report on a series of patients with cancer who also had MSC, which is an infrequent occurrence even at large cancer centers. Indeed, patients with MSC accounted for < 4% of all patients with candidemia treated at our institution between 1988 and 1993.10 Studies of unselected general patient populations with candidemia reported MSC rates ranging from 2% to 10%.8–19 The frequency of MSC does not appear to be higher in cohorts of candidemic patients who are immunosuppressed.10, 14, 16 However, because different Candida species have similar colonial morphologic appearances in routine culture media,23 the frequency of MSC may be underestimated. In fact, the results of a study conducted by Ahmad et al.,25 who used a more sensitive molecular technique (seminested polymerase chain reaction), suggest that the frequency of MSC may be higher than previously reported. Other techniques may also be useful in separating different Candida species in a single clinical specimen. Using the chromogenic culture medium CHROMagar (CHROMagar Co., Paris, France), Odds et al.26 found that the different colony colors and forms were easy to recognize in identifying different yeast species isolated from a clinical specimen.

Patients with C. albicans candidemia were selected as a control group because C. albicans continued to be the most common Candida species causing candidemia in our patient population during the study period.10, 27, 28 In the current study, MSC occurred predominantly in patients who had hematologic malignancies, and especially acute leukemia, as their underlying disease. Multivariate analysis identified leukemia as an independent factor predicting MSC, but not C. albicans candidemia. Accordingly, patients with MSC had higher severity-of-illness scores at the onset of infection and more protracted candidemia and were more frequently admitted to the ICU. In addition, the majority of patients developed MSC despite the use of antifungal prophylaxis. Taken together, these factors suggest that MSC is a marker of seriously compromised host defenses and probably a marker of severe mucosal surface damage due to cytotoxic chemotherapy.

Not surprisingly, persistent neutropenia was found to be a strong prognostic factor for poor response to antifungal therapy. Previous risk stratification models of candidemia in patients with cancer described the unfavorable prognostic significance of the presence of leukemia, persistent neutropenia, and high baseline APACHE II scores.19, 29

Most patients with MSC who received single-agent antifungal therapy responded suboptimally (35%) when compared with the C. albicans cohort (69%; P = 0.004). In contrast, the subset of patients with MSC who received combination therapy at the onset of infection appeared to have better outcomes, as three of these four patients experienced responses. Further studies are needed to address whether combination antifungal therapy is a preferable strategy for patients with cancer who develop MSC.

The current study had several limitations, due to its limited cohort size and retrospective nature, which prevented the assessment of other potentially important differences between the two cohorts (including differences pertaining to the presence of mucositis, dissemination, underlying disease activity, the impact of previous Candida colonization) and also did not allow the precise determination of MSC-attributable mortality.

In conclusion, MSC, although it is less common than candidemia caused by a single Candida species, is associated with severe morbidity and mortality. MSC is found primarily in neutropenic patients with underlying hematologic malignancies and occurs more frequently in these patients despite previous antifungal prophylaxis. The observed benefits of combination therapy in a subset of patients with MSC, although promising, should be examined further.

Ancillary