Efficacy and safety of amphotericin B formulations: a network meta-analysis and a multicriteria decision analysis

This document updates and expands the initial Infectious Diseases Society of America (IDSA) Fever and Neutropenia Guideline that was published in 1997 and first updated in 2002. It is intended as a guide for the use of antimicrobial agents in managing patients with cancer who experience chemotherapy-induced fever and neutropenia. Recent advances in antimicrobial drug development and technology, clinical trial results, and extensive clinical experience have informed the approaches and recommendations herein. Because the previous iteration of this guideline in 2002, we have a developed a clearer definition of which populations of patients with cancer may benefit most from antibiotic, antifungal, and antiviral prophylaxis. Furthermore, categorizing neutropenic patients as being at high risk or low risk for infection according to presenting signs and symptoms, underlying cancer, type of therapy, and medical comorbidities has become essential to the treatment algorithm. Risk stratification is a recommended starting point for managing patients with fever and neutropenia. In addition, earlier detection of invasive fungal infections has led to debate regarding optimal use of empirical or preemptive antifungal therapy, although algorithms are still evolving. What has not changed is the indication for immediate empirical antibiotic therapy. It remains true that all patients who present with fever and neutropenia should be treated swiftly and broadly with antibiotics to treat both gram-positive and gram-negative pathogens. Finally, we note that all Panel members are from institutions in the United States or Canada; thus, these guidelines were developed in the context of North American practices. Some recommendations may not be as applicable outside of North America, in areas where differences in available antibiotics, in the predominant pathogens, and/or in health care-associated economic conditions exist. Regardless of venue, clinical vigilance and immediate treatment are the universal keys to managing neutropenic patients with fever and/or infection.

Because of the increasing prevalence and changing microbiological spectrum of invasive fungal infections, some form of amphotericin B still provides the most reliable and broad spectrum therapeutic alternative. However, the use of amphotericin B deoxycholate is accompanied by dose-limited toxicities, most importantly, infusion-related reactions and nephrotoxicity. In an attempt to improve the therapeutic index of amphotericin B, three lipid-associated formulations were developed, including amphotericin B lipid complex (ABLC), liposomal amphotericin B (L-AmB), and amphotericin B colloidal dispersion (ABCD). The lipid composition of all three of these preparations differs considerably and contributes to substantially different pharmacokinetic parameters. ABLC is the largest of the lipid preparations. Because of its size, it is taken up rapidly by macrophages and becomes sequestered in tissues of the mononuclear phagocyte system such as the liver and spleen. Consequently, compared with the conventional formulation, it has lower circulating amphotericin B serum concentrations, reflected in a marked increase in volume of distribution and clearance. Lung levels are considerably higher than those achieved with other lipid-associated preparations. The recommended therapeutic dose of ABLC is 5 mg/kg/day. Because of its small size and negative charge, L-AmB avoids substantial recognition and uptake by the mononuclear phagocyte system. Therefore, a single dose of L-AmB results in a much higher peak plasma level (Cmax) than conventional amphotericin B deoxycholate and a much larger area under the concentration-time curve. Tissue concentrations in patients receiving L-AmB tend to be highest in the liver and spleen and much lower in kidneys and lung. Recommended therapeutic dosages are 3-6 mg/kg/day. After intravenous infusion, ABCD complexes remain largely intact and are rapidly removed from the circulation by cells of the macrophage phagocyte system. On a milligram-to-milligram basis, the Cmax achieved is lower than that attained by conventional amphotericin B, although the larger doses of ABCD that are administered produce an absolute level that is similar to amphotericin B. ABCD exhibits dose-limiting, infusion-related toxicities; consequently, the administered dosages should not exceed 3-4 mg/kg/day. The few comparative clinical trials that have been completed with the lipid-associated formulations have not demonstrated important clinical differences among these agents and amphotericin B for efficacy, although there are significant safety benefits of the lipid products. Furthermore, only one published trial has ever compared one lipid product against another for any indication. The results of these trials are particularly difficult to interpret because of major heterogeneities in study design, disease definitions, drug dosages, differences in clinical and microbiological endpoints as well as specific outcomes examined. Nevertheless, it is possible to derive some general conclusions given the available data. The most commonly studied syndrome has been empiric therapy for febrile neutropenic patients, where the lipid-associated preparations did not appear to provide a survival benefit over conventional amphotericin B deoxycholate, but did offer a significant advantage for the prevention of various breakthrough invasive fungal infections. For treatment of documented invasive fungal infections that usually involved hematological malignancy patients, no individual randomized trial has demonstrated a mortality benefit due to therapy with one of the lipid formulations. Results from meta-analyses have been contradictory, with one demonstrating a mortality benefit from all-cause mortality and one that did not demonstrate a mortality benefit. In the only published study to examine HIV-infected patients with disseminated histoplasmosis, clinical success and mortality were significantly better with L-AmB compared with amphotericin B deoxycholate; there were no differences in microbiological outcomes between treatment groups. The lipid-associated preparations were not significantly better than amphotericin B deoxycholate for treatment of AIDS-associated acute cryptococcal meningitis for either clinical or microbiological outcomes that were studied. In all of the trials that specifically examined renal toxicity, the lipid-associated formulations were significantly less nephrotoxic than amphotericin B deoxycholate. Infusion-related reactions occurred less frequently with L-AmB when compared with amphotericin B deoxycholate; however, ABCD had equivalent or more frequent infusion-related reactions than conventional amphotericin B, and this resulted in the cessation of at least one clinical trial. At the present time, this particular lipid formulation is no longer commercially available. For the treatment of most invasive fungal infections, an amphotericin B lipid formulation provides a safer alternative than conventional amphotericin B, with at least equivalent efficacy. As the cost of therapy with these agents continues to decline, these drugs will likely maintain their important role in the antifungal drug armamentarium because of their efficacy and improved safety profile.