Fluconazole for empiric antifungal therapy in cancer patients with fever and neutropenia.

BACKGROUND: Several clinical trials have demonstrated the efficacy of fluconazole as empiric antifungal therapy in cancer patients with fever and neutropenia. Our objective was to assess the frequency and resource utilization associated with treatment failure in cancer patients given empiric fluconazole antifungal therapy in routine inpatient care.
METHODS: We performed a retrospective cohort study of cancer patients treated with oral or intravenous fluconazole between 7/97 and 6/01 in a tertiary care hospital. The final study cohort included cancer patients with neutropenia (an absolute neutrophil count below 500 cells/mm3) and fever (a temperature above 38 degrees C or 100.4 degrees F), who were receiving at least 96 hours of parenteral antibacterial therapy prior to initiating fluconazole. Patients' responses to empiric therapy were assessed by reviewing patient charts.
RESULTS: Among 103 cancer admissions with fever and neutropenia, treatment failure after initiating empiric fluconazole antifungal therapy occurred in 41% (95% confidence interval (CI) 31%-50%) of admissions. Patients with a diagnosis of hematological malignancy had increased risk of treatment failure (OR = 4.6, 95% CI 1.5-14.8). When treatment failure occurred the mean adjusted increases in length of stay and total costs were 7.4 days (95% CI 3.3-11.5) and $18,925 (95% CI 3,289-34,563), respectively.
CONCLUSION: Treatment failure occurred in more than one-third of neutropenic cancer patients on fluconazole as empiric antifungal treatment for fever in routine clinical treatment. The increase in costs when treatment failure occurs is substantial.

Background

Immunocompromised patients are at risk of contracting serious fungal infections which cause significant morbidity and mortality [1-3]. The diagnosis of fungal infection in immunocompromised cancer patients is difficult for the clinician while the risk of fungal infection is high in patients with prolonged fever and neutropenia who do not receive antifungal therapy [4,5]. Empiric antifungal therapy for patients being treated with broad-spectrum antibiotics who have prolonged fever and neutropenia has been demonstrated to be efficacious, and it has become the standard of practice [5-7]. Newer antifungal agents have used for empiric antifungal therapy included liposomal amphotericin B [8] and caspofungin [9].

Fluconazole is active against the major fungal pathogens (except for mould) in neutropenic cancer patients although conventional amphotericin B is commonly used as a first-line agent for documented infections. Clinical trials have shown that fluconazole is an equally effective but less toxic alternative to amphotericin B for empiric antifungal therapy in cancer patients with prolonged fever and neutropenia [10-13]. These trials also evaluated clinical response to empiric antifungal therapy and found no significant difference in treatment failure with the rates of 25%–44% in fluconazole group compared to 33%–54% in patients on amphotericin B [10,11,13]. The relationships between treatment failure from empiric antifungal therapy and patient characteristics have been explored in a trial [11], although economic outcomes were not reported.

Relatively little information has been available regarding the frequency and resource use of treatment failure using fluconazole as empiric antifungal agent in cancer patients with fever and neutropenia in routine hospital inpatient care. To address these issues, we performed a retrospective cohort study with the goals of: 1) characterizing the patterns of care for cancer patients treated empirically with fluconazole; 2) identifying the rate of treatment failure; 3) assessing the increase in resource utilization associated with treatment failure; and 4) defining the risk factors for treatment failure.

Methods

Study site

The study was conducted in a 700-bed tertiary care teaching hospital, Brigham and Women's Hospital, Boston, MA, and has approximately 35,000 admissions per year. Patients were initially identified using the Brigham Integrated Computer System (BICS), which stored patient demographics (age, gender, race, insurance type, etc.) and medication orders. The study was approved by the Partners Human Research Committee.

Study population

The unit of analysis was the admission. All adult patients admitted and discharged between 7/1/1997 and 6/30/2001, who received oral or intravenous fluconazole were eligible. Cancer patients with fever and neutropenia, and who underwent empiric fluconazole antifungal therapy were included in the final study cohort if they met all of the following inclusion criteria: (1) an oncologic diagnosis; (2) an absolute neutrophil count (ANC) < 500 cells/mm3 within 4 days prior to starting fluconazole; (3) receiving at least 96 hours of parenteral systemic antibacterial therapy; (4) a fever (a temperature > 38°C or 100.4°F) within 4 days prior to start of fluconazole and (5) no documented invasive fungal infections as the indication for initiation of fluconazole (based on the judgment of the care providers as reported in the patients medical record). The second and fourth criteria were referred to any day within the 4 days prior to starting fluconazole. For patients with multiple courses of therapy with fluconazole during the same admission, only those who had neutropenia and fever prior to the first course of fluconazole therapy were included in the final study cohort.

Data

For patients eligible for the study, data were obtained from both the BICS and the Transition Systems Incorporated (TSI; Boston, MA). For the final study cohort, additional data were collected through a chart review. Demographic and other data including age, gender, race, insurance status (classified as fee-for-service, health maintenance organization, Medicare, Medicaid and none), course of therapy on fluconazole, type of hospital service, primary diagnosis, discharge diagnosis-related group (DRG) and laboratory results including neutrophil count, bacterial and fungal cultures were collected from BICS. Hospital costs were downloaded from the TSI system, which uses a process costing technique to estimate the expenses incurred by cost centers in providing products and services.

In addition, charts were reviewed on the final study cohort to collect additional information for a variety of variables. These included the presence of major underlying diseases, Charlson comorbidity score [14], the presence of organ transplantation, date with fever, information about the use of immunosuppressive therapy before empiric fluconazole therapy, dosage and duration of fluconazole, and presence of antifungal therapy before initiation and after end of the empiric fluconazole therapy. Indications for stopping the empiric fluconazole antifungal therapy were classified as: positive fungal culture one day after starting fluconazole up to the end of course without changing to negative culture, persistent fever at the end of course with negative bacterial and fungal cultures throughout the course of fluconazole, ADE and/or toxicity due to fluconazole, patient death, care withdrawn and planned course of fluconazole completed defined by fluconazole orders having been completed and patients no longer having fever and neutropenia. Information was also collected on whether the patient was discharged on empiric fluconazole therapy or readmitted after being discharged.

Outcome measures

The primary outcome was treatment failure, which was considered present if any of the following criteria were met: development or persistence of fungal infection evidenced by positive fungal culture during empiric fluconazole therapy and/or up to 7 days after fluconazole was stopped, persistence of fever during empiric fluconazole therapy and/or up to 7 days after fluconaozle therapy without the presence of positive bacterial and fungal cultures, ADE and/or toxicity due to fluconazole, and death from any cause within 7 days after fluconazole was stopped. Positive bacterial and fungal cultures were identified from laboratory reports of tests on specimens from blood, bronchial alveolar lavage fluid, sputum, stool, urine and catheter tips. The date of the laboratory report was used to categorize patients with positive fungal cultures before and after the initiation of empiric therapy. The final study cohort included patients who had positive fungal cultures reported before the initiation of fluconazole, but who did not meet the criteria for diagnosis of invasive fungal infections (no patients had proven or probable invasive fungal infection while one patient had possible invasive fungal infection) as defined by the European Organization for Research and Treatment of Cancer and the National Institute of Allergy and Infectious Diseases Mycoses Study Group [1]. An ADE was defined as "an injury resulting from an intervention related to a drug" [15]. Length of stay (LOS) was measured in days before and after empiric fluconazole antifungal therapy was initiated. Total hospital costs were calculated using a discount rate of 3% annually with the year 2001 as reference. In-hospital mortality was also measured.

Analyses

For each of the patient characteristics, dosage and duration of fluconazole, and outcomes of measurement, counts with proportions, means with standard deviations, and medians with ranges were determined as appropriate. Univariate comparisons between admissions in which treatment failure developed and those in which it did not were made using the Wilcoxon rank-sum statistic for non-normal continuous variables and the chi-square statistic or Fisher's exact test for categorical variables. To identify independent correlates of treatment failure, multivariate analyses was performed using logistic regression. All variables prior to empiric fluconazole antifungal therapy were included in the baseline model if they had an association with the outcome with a P value less than 0.10 in the univariate analysis. Stepwise logistic regression was used to identify correlates of treatment failure, and variables were retained using a threshold of P < 0.05. Multivariate comparisons were made using generalized linear modeling, with length of stay and costs as the dependent variable and treatment failure as the main covariate of interest, adjusting for age, gender, insurance status, length of stay before empiric fluconazole therapy, and DRG weight.

A sensitivity analysis was conducted using two approaches. The first approach derived 95 percent confidence intervals using bootstrap estimates (1,000 repetitions) to develop a statistical range for the estimates of total costs and length of stay. With the bootstrap method, one can also determine the stability of the models' effect estimates [16]. The second approach repeated resource utilization models after excluding patients who had positive fungal cultures at baseline because these patients might not be necessarily categorized as on empiric antifungal therapy although none of them had proven invasive fungal infections. Patients who had fungal infections at baseline could have higher resource utilization than others.

The final regression models to predict length of stay and total costs were checked for outliers and influential points [17]. Outliers were identified by examining the jackknife residuals for all patients with studentized residuals greater than 2 in absolute magnitude; six outliers were found. No influential point was detected; the largest value of Cook's distance was 0.18. We therefore present analyses without excluding outliers. All analyses were performed using the SAS statistical package [18].

Results

Among the total of 1,537 cancer patient admissions in which fluconazole was administered orally and/or intravenously during their hospitalization in the four year study period, 103 (6.7%, 95% CI 5.5% – 8.0%) were treated with empiric fluconazole antifungal therapy for neutropenia and fever after receiving at least 96 hours of parenteral systemic antibacterial therapy. Among these 103 patient admissions, empiric antifungal treatment failed in 42 (40.8%, 95% CI 31.3% – 50.3%) admissions (Table 1). The most frequent causes of treatment failure were positive fungal cultures (20 patients) and persistence of fever with negative bacterial and fungal cultures (19 patients) during empiric fluconazole treatment or within 7 days after the end of empiric therapy. Other causes included fluconazole-associated hepatotoxicity (n = 1) and death earlier than 7 days after the end of empiric fluconazole therapy (n = 5). Three patients had multiple causes of treatment failure. Among the 20 patients with positive fungal cultures after the initiation of empiric therapy, 7 had positive fungal cultures from the same type of specimen prior to beginning empiric therapy.

Table 1

Measure of Treatment Failure

Variable, No. (%)	Alln = 103
Treatment failure	42 (40.8)
Reason for treatment failure *
Positive fungal cultures **	20 (19.4)
During empiric fluconazole therapy	16 (15.5)
From day 1 to 7 after fluconazole stopped	6 (5.8)
Persistent fever with negative bacterial and fungal cultures †	19 (18.5)
During empiric fluconazole therapy	17 (16.5)
From day 1 to 7 after fluconazole stopped	4 (3.9)
ADE/toxicity due to fluconazole	1 (1.0)
Death within 7 days after fluconazole stopped	5 (4.9)

ADE, adverse drug event; * Three patients had multiple major reasons including 1 patient had positive fungal culture within 7 days after fluconazole stopped and persistent fever with negative bacterial and fungal cultures during empiric therapy, and 2 patients had persistent fever after initiation of fluconazole and died within 7 days after the end of fluconazole therapy; ** Positive fungal cultures referred to positive tests on specimen from blood, bronchial alverolar lavage, sputum, stool, urine and catheter tips and no report of negative results for repeated test on specimen from the same site, † fever present at least on the last day of fluconazole course and/or on day 7 after fluconazole stopped.

ADE case report

One treatment failure was due to a serious ADE in a 28-year-old woman with recurrent Hodgkin's disease who had been admitted to the hospital for high dose chemotherapy and bone marrow transplantation. Antifungal treatment was started with intravenous fluconazole 200 mg daily on hospital day 9. On day 12, values of liver function test were elevated (ALT = 147 IU/L, AST = 205 IU/L and BILT = 2.0 mg/dL). Fluconazole was discontinued on day 13. The liver function results returned to a normal value on day 15 and remained normal until the day of discharge. The patient did have chronic active hepatitis B and was on lamivudine but the dose was unchanged during the hospital course.

Characteristics

The final cancer patient cohort on empiric fluconazole antifungal therapy (Table 2) had a mean age of 51 (+/- 14) years, and was 60% female and 95% white. All patients were receiving care from the oncology or bone marrow transplant services prior to fluconazole treatment. Over three quarters of the patients (76.7%) had hematological malignancies including leukemia and myelodysplastic syndrome (45.6%), lymphoma (22.3%), myeloma (4.9%) and Hodgkin's disease (3.9%), while the remainder of the patients (23.3%) had solid tumors including breast cancer (20.4%), endometrial cancer, neuroendocrine cancer and ovarian cancer. The vast majority of patients (96.1%) were on immunosuppressive therapy with steroids, cyclosporine or chemotherapy and more than half had undergone bone marrow transplantation before fluconazole treatment.

Table 2

Patient Characteristics

	All(n = 103)	Treatment Failure(n = 42)	No Treatment Failure(n = 61)
Age, year, mean (SD)	51 (14.1)	51 (14.3)	51 (14.0)
Female, No. (%)	62 (60.2)	24 (57.1)	38 (62.3)
Nonwhite, No. (%)	5 (4.9)	4 (9.5)	1 (1.6)
Uninsured or Medicaid, No. (%)	12 (11.7)	6 (14.3)	6 (9.8)
Primary diagnosis, No. (%)
Hematological malignancies	79 (76.7)	38 (90.5)	41 (67.2) ‡
Acute leukemia/MDS	40 (38.8)	20 (47.6)	20 (32.8)
Chronic leukemia	7 (6.8)	3 (7.1)	4 (6.6)
Lymphoma	23 (22.3)	11 (26.2)	12 (19.7)
Myeloma	5 (4.9)	3 (7.1)	2 (3.3)
Hodgkin's disease	4 (3.9)	1 (2.4)	3 (4.9)
Solid tumors	24 (23.3)	4 (9.5)	20 (32.8)
Underlying disease, No. (%)	11 (10.7)	5 (11.9)	6 (9.8)
Cardiovascular disease	4 (3.9)	2 (4.8)	2 (3.3)
Others *	7 (6.8)	3 (7.1)	4 (6.6)
DRG weight, mean (SD)	7.4 (3.7)	7.0 (3.6)	7.7 (3.7)
Charlson score, mean (SD)	3.0 (1.7)	2.6 (1.4)	3.3 (1.8) ‡
Charlson score ≤ 4, No. (%)	80 (77.7)	38 (90.5)	42 (68.9) ‡
Treatment before fluconazole started, No. (%)
Steroid	17 (16.5)	7 (16.7)	10 (16.4)
Chemotherapy	98 (95.2)	38 (90.5)	60 (98.4)
Cyclosporin	6 (5.8)	3 (7.1)	3 (4.9)
History of Transplant, No. (%)
Bone marrow transplant	56 (54.4)	23 (54.7)	33 (54.1)
Lab results before fluconazole started, No. (%)
ANC ≤ 100/mm3	74 (71.8)	33 (78.6)	41 (67.2)
Positive fungal cultures **	19 (18.5)	9 (21.4)	10 (16.4)
Candida albicans	15 (14.6)	8 (19.1)	7 (11.5)
Yeast species	7 (6.8)	3 (7.1)	4 (6.6)
Duration before fluconazole started, day, mean (SD)
Fever	7.9 (6.8)	7.5 (6.9)	8.3 (6.8)
Neutropenia	6.5 (6.9)	6.5 (6.8)	6.4 (7.0)
Deceased, No. (%)	8 (7.8)	6 (14.3)	2 (3.3) †

MDS, myelodysplastic syndrome; SD, standard deviation; DRG, diagnosis-related group; ANC, absolute neutrophil count;

* Other underlying diseases included chronic pulmonary disease, connective tissue disease, ulcer disease, mild liver disease, mild to moderate diabetes;

** Positive fungal cultures were reported for tests on specimen collected from sputum, stool and urine; three patients had multiple positive cultures from different type of specimen;

† 0.05 ≤ P < 0.10, ‡ 0.001

† 0.05 ≤ P < 0.10, ‡ 0.001