Amphotericin B Lipid Complex in the Treatment of Invasive Fungal Infections: Results of the Collaborative Exchange of Antifungal Research (CLEAR), an Industry-Supported Patient Registry
Year: 2005
Type: article
Study design: Observational Study
Abstract: Many factors confound the conduct, reproducibility, and interpretation of prospective therapeutic trials of antifungal agents. In general, clinical trials of antifungal agents involve relatively small numbers of patients, owing to the relative rarity of certain mycoses, strict eligibility criteria, and the complexity of the comorbid conditions encountered among patients who develop invasive fungal infections. Compounding these problems is the relative lack of diagnostic methods that permit accurate identification of potential study candidates early enough during the course of the infection to facilitate study enrollment. In recognition of these substantial obstacles, some investigators have advocated the use of a more innovative approach to the study of invasive mycoses—for example, the use of surrogate markers and nonculture techniques to identify potential study candidates and to assess response [1]. Finally, despite the growing size of the population of at-risk patients and the recognition of new and emerging fungal pathogens, the study of the diagnosis and treatment of invasive fungal infections is underfunded at all levels. The Collaborative Exchange of Antifungal Research (CLEAR) database was developed in the mid-1990s in response to some of these unresolved issues. The database has enhanced our understanding of the epidemiology of invasive fungal infections and clinical experience with amphotericin B lipid complex (ABLC) injection (Enzon Pharmaceuticals). ABLC is a composition of amphotericin B and 2 phospholipids (dimyristoylphosphatidylcholine and dimyristoylphosphatidylglycerol) in a 1 : 1 drug-to-lipid molar ratio [2]. ABLC was developed to reduce the nephrotoxicity associated with the conventional formulation of amphotericin B deoxycholate (AmB), thus allowing for administration of higher doses of ABLC than AmB for longer therapeutic courses [3]. Although conventional AmB has been considered as the reference standard of antifungal therapy for many invasive mycoses, the high incidence of renal toxicity and other side effects associated with its use have resulted in a recent reevaluation of its position as first-line therapy for many mycoses [4–7]. ABLC is rapidly absorbed by the reticuloendothelial system, resulting in a lower concentration of amphotericin B in the plasma and kidneys and a greater distribution of active drug at tissue sites. Therefore, ABLC may be beneficial in the treatment of fungal infections when renal dysfunction is a concern or as second-line therapy when AmB-induced nephrotoxicity occurs. The CLEAR database provides data on efficacy and renal safety for 3514 patients treated with ABLC from 1996 to 2000 at >160 institutions in the United States and Canada. One of the strengths of the CLEAR database is that it reflects a broad spectrum of patients treated for invasive fungal infections in North America and avoids the risk of excluding particular groups of patients, which often occurs during recruitment of patients for phase 3 clinical trials. Similar to other observational and retrospective studies of antifungal therapy, the CLEAR database has its distinct limitations: it was retrospective, the data collection method was based on voluntary reporting and, thus, was subject to potential enrollment bias, objectively defined response criteria were lacking, and follow-up was limited to completion of therapy with ABLC or discharge from the hospital, whichever was sooner. However, this database has value because it provides a unique opportunity to examine various aspects of therapy with ABLC among a broad scope of patients with various invasive fungal infections. Table 1 depicts the demographic characteristics of the entire population of patients enrolled in CLEAR. The strength of these data and the conclusions that can be drawn from the database are based on the large number of patients enrolled, the varying underlying conditions studied, and the range of fungal pathogens treated. These pathogens include not only Candida and Aspergillus species but also new and emerging pathogens, such as non-Aspergillus moulds (e.g., Fusarium species and the Zygomycetes), the endemic dimorphic fungi (e.g., Histoplasma capsulatum and Coccidioides immitis), and Cryptococcus species. Patients included in CLEAR are stratified according to unresponsiveness to or intolerance of prior antifungal agents and according to whether they received ABLC as first-line or second-line therapy. The patient-specific data collected in CLEAR allow for clinical assessment based on the presence or absence of significant underlying disease, thus permitting the investigator to examine the relationship between underlying disorders and clinical outcomes. Demographic characteristics of all patients registered in the Collaborative Exchange of Antifungal Research (CLEAR) database (N = 3514). The CLEAR database was designed and governed by a steering committee of physicians and scientists from across the United States and Canada who have made substantial contributions to the study of invasive fungal infections (see the list of Steering Committee members at the end of the text). This group was instrumental in determining the most relevant information to include in the database, as well as the data collection techniques. Patient inclusion criteria. To be eligible for enrollment in the CLEAR database, patients must have received (in the hospital setting) at least 4 doses of ABLC for treatment of a proven or suspected fungal infection. Patients were prospectively enrolled in the CLEAR program. On the basis of pre–Health Insurance Portability and Accountability regulations, patient charts were retrospectively reviewed, and data were collected on standardized case report forms after identifying information was removed. During the study period, institutional review board consent for retrospective collection of data was obtained at each study site. Data collection. A standardized case report form was used to collect and record patient data, including demographic characteristics, infection characteristics, diagnostic methods, reason for initiation of therapy with ABLC, dose and duration of therapy with ABLC and prior or concomitant medications, and clinical outcomes. Specific underlying medical conditions were noted on the case report form, and ⩾1 underlying condition could be documented for each patient. Information regarding the specific fungal infection and sites of involvement was recorded. Patients were evaluated by investigators at the end of therapy with ABLC or at hospital discharge, if outpatient therapy was to be continued. The drug manufacturer at the time (The Liposome Company) provided funding for case management and submission of the information to the registry. Mortality data were not collected for the entire length of the study and thus were not analyzed in CLEAR. Data on infusion-related reactions were not collected. In addition, data on the use of surgical debridement as primary or adjuvant treatment were not collected. Investigators were asked to classify the nature of the fungal infection, the reason for initiating therapy with ABLC, and the outcome of treatment (table 2). Fungal infections were classified as proven or suspected. Clinical evaluation and methods used to diagnose a fungal infection included clinical signs and symptoms, radiographic evidence, blood and tissue cultures, serology, histology, and sputum collection. Although there were no strict diagnostic criteria set forth for assessing proven or suspected infection, investigators were encouraged to use guidelines similar to those later adopted by the European Organization for Research and Treatment of Cancer and the Mycoses Study Group of the National Institute for Allergy and Infectious Diseases [8]. Collaborative Exchange of Antifungal Research (CLEAR) classifications. Investigators classified the rationale for therapy with ABLC according to ⩾1 of the following 4 categories: (1) infection was refractory to prior antifungal therapy; (2) the patient had underlying renal disease; (3) the patient was intolerant of prior antifungal therapy, usually determined on the basis of infusion-related toxicity, an increasing serum creatinine level, or hepatotoxicity; or (4) the patient had no underlying renal disease and no prior antifungal treatment. After the data were collected, the underlying renal disease classification was further broken down on the basis of prior exposure to antifungal therapy. This additional categorization enabled analyses of patients who received ABLC as first-line therapy rather than second-line therapy. Patients who received ABLC as second-line therapy had previously experienced therapy failure with other antifungal drugs, primarily fluconazole, conventional AmB, or itraconazole. At the time this study was conducted, voriconazole and the echinocandins were not yet approved for clinical use. Clinical outcomes were recorded on the basis of each investigator's clinical assessment and were consistent with the following guidelines: cured (i.e., clinical signs and symptoms of fungal infection were absent), improved (i.e., clinical signs and symptoms of fungal infection had improved), stable (i.e., there was no change in overall clinical findings, compared with previous reporting period), deteriorated (i.e., clinical signs and symptoms of fungal infection had worsened, including death if recorded as such), and indeterminate (i.e., clinical signs and symptoms of fungal infection were insufficient to make an evaluation). Patients in the indeterminate category were not considered to be eligible for evaluation, because of lack of information, and were excluded from the analyses. Stable disease, which was defined as no worsening of the disease, was considered to be an acceptable outcome in CLEAR. For purposes of the analyses in this supplement, certain authors have chosen to consider only cured and improved responses as indicative of a successful response to therapy, to more clearly reflect clinical assessments set forth in current antifungal clinical trials. However, stable responses may provide important insights into treatment strategies for difficult-to-treat infections and are reported in this supplement, where appropriate, for reader consideration. The CLEAR program was neither structured nor intended to serve as a comparative study; thus, the authors of this publication did not compare efficacy results in CLEAR with those of other antifungal agents or studies. Data analysis. The total (or cumulative) ABLC dose was calculated for each patient by summing all doses of ABLC administered between the start and end dates of therapy. The actual duration of ABLC dosing (in days) was calculated for each individual dose record as follows: (last dose date - first dose date) + 1. The average daily dose was calculated as follows: (total dose/weight)/actual duration. The peak daily dose of ABLC was determined by identifying the largest dose of ABLC given on a particular day. Analyses of the renal safety of ABLC included evaluation of baseline and end-of-treatment serum creatinine levels, change from baseline serum creatinine levels, the occurrence of doubling in serum creatinine from baseline levels by the end of treatment, and the need for any new dialysis. In addition, predicted creatinine clearance at baseline and end of therapy, as well as change from baseline, was calculated on the basis of the Cockroft-Gault formulas: for male patients, weight in kg × (140 - age)/(72 × serum creatinine), and for female patients, weight in kg × (140 - age)/(72 × serum creatinine) × 0.85. In analyzing the reasons for starting therapy with ABLC, priority was given to the category chosen initially, such that if >1 category was chosen for a patient, the patient was categorized under the first checked category in the order listed: (1) refractory to prior antifungal therapy, (2) underlying renal disease, (3) intolerant of prior antifungal therapy, and (4) no underlying renal disease and no prior antifungal treatment. To facilitate analysis of the data on the basis of underlying medical conditions and sites of infection, the parameters were placed in hierarchical order of importance on the basis of the pathogen under scrutiny. Each separate analysis in this supplement lists the hierarchy by which underlying conditions and sites of infection were categorized. When statistical analyses were performed, the significance of difference in proportions was determined by use of χ2 or Fisher's exact test, as appropriate. The difference in continuous variables was detected by use of a nonparametric median test. All P values were 2-sided, and P < .05 was considered to be statistically significant. For change in serum creatinine level from baseline and change in predicted creatinine clearance, no assumptions of the underlying distribution were made, and the nonparametric median test was used to compare the difference between the cohorts. All data processing and statistical analyses were performed by use of SAS System (version 8.2; SAS Institute) on a Windows 2000 or XP operating system. Multivariate analyses were not reported in manuscripts dealing with clinical efficacy. Clinical response rates were investigated in populations that were already heavily narrowed down to include specific pathogens or species, specific sites or type of sites of infection, and neutrophil status. The numbers of patients in each category would not have allowed for a multivariate analysis. The goal with each of these studies was achieved with the analyses provided. Over the course of the last 5 years, information from the CLEAR database has been published in scientific journals and has been presented at numerous professional meetings. In this publication, the contributing authors have elected to report on the efficacy and renal safety of ABLC for the treatment of the fungal infections that are of greatest concern in our current clinical environment: invasive candidiasis, invasive aspergillosis, infections due to non-Aspergillus moulds (e.g., fusariosis and zygomycosis), endemic mycoses, and cryptococcosis. In addition, an analysis of the renal safety of ABLC within the entire CLEAR population of 3514 patients is presented. Importantly, the results drawn from this database confirm the observations of other published studies of therapy with ABLC and provide data on outcomes for a significant number of patients with infections caused by less-common emerging pathogens that have not been the subject of rigorous clinical study. Thus, this publication represents our attempt to distill and provide the most useful information from the CLEAR registry with regard to its renal safety profile and its efficacy against selected medically important fungal pathogens. Members of the CLEAR Steering Committee and their affiliations at program inception are as follows: Elias J. Anaissie (Arkansas Cancer Research Center, Little Rock), Larry M. Baddour (Mayo Clinic, Rochester, MN), Eric J. Bow (University of Manitoba, Winnipeg, Canada), James A. Boyle (formerly of The Liposome Company, Princeton, NJ), Steven Gelone (Temple University School of Pharmacy, Philadelphia, PA), Roger H. Herzig (University of Louisville Hospital, Louisville, KY), John W. Hiemenz (Medical College of Georgia, Augusta), James I. Ito, Jr. (City of Hope National Medical Center, Duarte, CA), Bernard M. Kubak (University of California, Los Angeles), Peter K. Linden (University of Pittsburgh Medical Center, Pittsburgh, PA), John R. Perfect (Duke University Medical Center, Durham, NC), John H. Rex (formerly of University of Texas Medical School, Houston; currently with Astra-Zeneca Pharmaceuticals, Manchester, UK), Michael G. Rinaldi (University of Texas Health Science Center at San Antonio), Trudy N. Small (Memorial Sloan-Kettering Cancer Center, New York, NY), Thomas J. Walsh (National Cancer Institute, Bethesda, MD), Joseph M. Wiley (Sinai Hospital of Baltimore, Baltimore, MD), and John R. Wingard (University of Florida Medical Center, Gainesville). I acknowledge the contributions of D. Scott Batty in these analyses of the CLEAR database and the assistance of The Spellman Group in the editing of this manuscript. Financial support. Enzon Pharmaceuticals. Potential conflicts of interest. P.G.P.: member of the ad hoc advisory boards for and receives grant support from Enzon, Pfizer, Schering-Plough, Merck, and Vicuron; speakers' bureaus for Merck, Pfizer, and Fujisawa.
Source: Clinical Infectious Diseases
DOI: 10.1086/429329
Author: Peter G. Pappas
Institutions: University of Alabama at Birmingham, UAB Medicine
Language: English
Subfield: Infectious Diseases
Field: Medicine
Domain: Health Sciences
Open Access status: bronze
OpenAlex ID: W2149642450