Supported by an unrestricted educational grant from Pfizer, Inc.
Volume II, Issue 2 - June 1999
Supported by an unrestricted educational grant from Pfizer, Inc.
Conventional amphotericin B is complexed with a lipid component, sodium deoxycholate, that enables solubility in aqueous solutions, thus allowing for parenteral administration. Over the past 2 decades, researchers have investigated the utility of incorporating amphotericin B into phospholipid vesicles (liposomes) and/or cholesterol esters in order to provide larger amounts of parent drug and concomitantly, less nephrotoxicity. To date, 3 lipid formulations of amphotericin B are commercially available: amphotericin B lipid complex (ABLC, Abelcet); amphotericin B cholesteryl sulfate complex, also called amphotericin B colloidal dispersion (ABCD, Amphotec); and liposomal amphotericin B (L-AmB, AmBisome) (Table 1).
Amphotericin B, a polyene antibiotic synthesized by Streptomyces nodosus, an organism found in soil, is capable of binding irreversibly to ergosterol in the fungal cytoplasmic membrane, thus increasing membrane permeability with ultimate fungal cell death. Lipid formulations of amphotericin B were developed to improve the amphipathic nature of amphotericin B and facilitate drug insertion within the fungal cytoplasmic membrane while reducing uptake in human cells, thereby limiting toxicity.
The pharmacokinetics and tissue distribution of the lipid amphotericin B formulations are largely influenced by the physical and chemical properties of the lipid vehicle. For example, the biologic half-life of liposome vesicles is increased with smaller size, positive surface electrical charge, and higher cholesterol concentration in the phospholipid bilayers. The chemical properties and physical characteristics of available amphotericin B products are outlined in (Table 2). Each lipid formulation is unique with respect to its lipid content, configuration, and molar content of amphotericin B. L-AmB is the only product that contains true liposomes.
The pharmacokinetics and pharmacodynamics of the lipid formulations of amphotericin B differ greatly from those of AmBD and each lipid formulation has distinct pharmacologic properties, including clearance, area under the curve (AUC), maximum plasma concentration (Cmax), and volume of distribution. When differences in units of clearance are considered, the clearance of all lipid-based formulations is much lower than that of AmBD. Infusion of L-AmB results in strikingly higher AUC and Cmax values than are achieved with AmBD or the other lipid formulations. Although the 3 lipid agents are distributed extensively into tissues, especially the primary reticuloendothelial organs (liver, spleen, and lungs), the apparent volume of distribution for ABCD and ABLC after single-dose infusion is substantially greater than that of L-AmB. For all 3 lipid-based drugs (ABCD, ABLC, and L-AmB) kidney tissue concentrations are substantially lower than those of AmBD. These disparities in pharmacologic properties may be linked to the binding affinity and subsequent release of amphotericin B by the individual lipid vehicles. The liposomal component of L-AmB has the tightest bond to amphotericin B, explaining in part its high plasma concentrations and lower clearance rates relative to the other lipid formulations. As a general principle, the lipid formulations of amphotericin B at dosages 5- to 10-fold higher than dosages of AmBD can be administered safely. Animal studies have demonstrated lipid-based amphotericin B preparations are clearly less toxic to mammalian cells than AmBD.
The majority of clinical efficacy data related to the lipid-based amphotericin B drugs are derived from compassionate use studies and small case series. Few randomized studies have been performed comparing the lipid formulations to AmBD or to each other. Table 3 provides the Food and Drug Administration (FDA) current approved indications for ABLC, ABCD, and L-AmB. This paper summarizes the available efficacy data for febrile neutropenia, invasive fungal infections, aspergillosis, and cryptococcosis. The adverse events associated with these agents will be discussed separately.
There have been 2 randomized, double-blind clinical trials comparing the efficacy of lipid formulations of amphotericin B (ABCD and L-AmB) to AmBD for empiric therapy in adult febrile neutropenic patients. In both studies inclusion criteria included neutropenia (<500/mm3), persistent fever despite empiric bacterial therapy, and lack of documented fungal infection or recent exposure to AmBD. Dosing regimens were as follows: ABCD 4 mg/kg/d, L-AmB 3mg/kg/d, and AmBD .06 to .08 mg/kg/d. In the ABCD versus AmBD trial, dosage adjustments were allowed for toxicity or lack of clinical response. Successful outcomes were defined as survival for >7 days after the last dose of study drug, resolution of fever, lack of suspected or documented fungal infection, and lack of drug discontinuation due to adverse events. A successful response was noted in 49 (50%) of 98 ABCD recipients and in 41 (43%) of 95 AmBD recipients (P = 0.31). In a larger, multicenter trial the composite success rate was similar for L-AmB (50% [172 of 343 patients]) and AmBD (49% [170 of 344 patients]). Fewer proven breakthrough fungal infections occurred in those treated with L-AmB (3.2% versus 7.8%, P = 0.009). In general, ABCD and L-AmB have comparable efficacy to AmBD for the empiric treatment of febrile, neutropenic patients but are significantly less likely to be associated with renal dysfunction.
Invasive fungal infections are an important cause of morbidity and mortality in immunocompromised patients. Here, results of 3 open label, noncomparative studies are briefly discussed. One trial, an expanded access study, evaluated the safety and efficacy of ABLC (5 mg/kg/d) in 556 cases of invasive fungal infections in patients refractory or intolerant to AmBD. Of the 291 patients with mycologically confirmed infection, 167 (57%) had a complete or partial response to ABLC, including 55 of 130 (42%) cases of aspergillosis, 28 of 42 (67%) cases of disseminated candidiasis, and 17 of 24 (71%) cases of zygomycosis. Patients with AIDS and bone marrow transplant recipients had significantly reduced response rates compared with the response rates among patients who had hematologic malignancies and solid organ transplants.
In a second trial, 168 patients with documented or suspected systemic mycoses refractory to AmBD received ABCD via an expanded access program. Dosing of ABCD was varied. Initial dosing was 0.5 to 4.0 mg/kg/d and could be increased up to 6 mg/kg/d. Complete clinical response or improvement was noted in 48 of 97 (49%) evaluable patients. Among the 32 patients infected with Aspergillus species, 11 (34%) responded and 19 (58%) of the 33 patients infected with Candida species responded. In a third multicenter, open label study, 126 immunocompromised patients who were refractory or intolerant to AmBD received L-AmB at dosages of 0.5 to 5.0 mg/kg/d. Clinical efficacy data were available for 108 patients, 52 of whom had infections due to Candida species and 34 due to Aspergillus species. The response rate was 58% in Aspergillus infections and 84% in Candida infections, with cure rates of 34% and 76%, respectively. These efficacy data obtained from compassionate use studies suggest the lipid formulations of amphotericin B are effective against invasive fungal infections in patients refractory or intolerant to AmBD. Response rates for all 3 lipid-based drugs are higher in patients with Candida infection compared to patients with Aspergillus infection.
In addition to the open label, compassionate use studies described above, a retrospective analysis compared ABCD and AmBD for the treatment of invasive aspergillosis. All patients had proven or probable Aspergillus infection. Dosing strategies were highly variable within the ABCD group (n = 82) who received 0.5 to 8.0 mg/kg/d and the AmBD group (n = 261) who received 0.1 to 1.4 mg/kg/d. The overall response rate was significantly greater in the ABCD group (48.8% versus 23.4%, P <0.001. However, these retrospective data are difficult to interpret given the disparities in dosing and marked demographic differences between the 2 study groups, as well as lack of a matched cohort design. For example, neutropenia was significantly more common in the AmBD group. Additionally, a prospective, randomized trial compared the efficacy of 2 dosages of L-AmB for invasive aspergillosis in neutropenic patients or bone marrow transplant recipients. Forty-one patients were randomized to receive 1 mg/kg/d and 46 patients to receive 4 mg/kg/d. The clinical response rate was better in the low-dose group, 64% versus 48%, respectively (P = 0.14) and radiographic response rates were similar (58% for 1 mg/kg/d and 54% for 4 mg/kg/d). Thus, in this important trial no advantage of the higher, more expensive dosage of L-AmB was observed. These somewhat conflicting results emphasize that additional, adequately powered, randomized, prospective studies are necessary to clarify the role and optimum dosage of lipid-based amphotericin drugs in the treatment of invasive aspergillosis.
Currently, L-AmB is the only lipid-based amphotericin drug approved by the FDA for treating cryptococcal infection. To date, 2 comparative studies have been performed. In a trial comparing ABLC to AmBD, 55 individuals infected with human immunodeficiency virus (HIV) and cryptococcal meningitis were randomized into 3 small cohorts to receive ABLC (1.2, 2.5, or 5.0 mg/kg/d) and in 1 cohort to receive AmBD (0.7 mg/kg/d). Initial therapy, given for 2 weeks, was followed by 4 weeks of ABLC at dosages of 2.5 or 5.0 mg/kg/d or AmBD at a dosage of 1.2 mg/kg thrice weekly. No significant differences in clinical, mycologic, or overall responses were noted between the 3 ABLC cohorts and the AmBD cohort but the ABLC groups had a disproportionate share of negative predictive outcome factors including persistently positive cerebrospinal fluid cultures. In a smaller study, 28 HIV-infected patients with cryptococcal meningitis received L-AmB (4 mg/kg/d) or AmBD (0.7 mg/kg/d) for 3 weeks followed by fluconazole (400 mg/d) for 7 weeks. The clinical response rates were similar between the 2 groups (87% with L-AmB and 83% with AmBD) but the L-AmB group experienced more rapid sterilization of their cerebrospinal fluid (median 14 days versus >21 days). Taken together, these 2 small studies support the use of L-AmB and ABLC as alternatives to AmBD for therapy of cryptococcal meningitis in patients who have preexisting renal dysfunction or develop worsening renal function on AmBD treatment. Limited human data and findings in animal model studies indicate ABCD may also be effective therapy for cryptococcosis.
(Table 4) estimates the frequency of adverse events attributed to lipid based formulations of amphotericin B, based on available data from both open label and comparative clinical trials as well as our own experience. Such estimates are imprecise because of many factors, including lack of comparability of dosing, variability in administration of premedications prior to dosing, no standardization of definitions of toxicity or adverse events, and incompleteness of data across studies. The primary adverse events are related to infusion (e.g., fever, chills, and nausea and vomiting) and nephrotoxicity. Infusion-related events connected with ABCD and ABLC appear to be considerably more common than those associated with L-AmB and similar in frequency to infusion-related events associated with AmBD. In 1 large comparative trial, during initial infusion without premedication, L-AmB recipients experienced significantly less fever and chills than AmBD recipients (fever, 16.9% versus 43.6%, respectively; and chills, 18.4% versus 54.4%, respectively). Fever and chills associated with the lipid-based drugs tend to occur mainly after the first 2 infusions and are less frequent with subsequent infusions. Infusion-related hypoxia has been documented in as many as 25% of ABCD and ABLC recipients but is usually reversible and without long-term sequelae. Infusion-related adverse events associated with the lipid-based drugs, just as with AmBD, can be attenuated or prevented by premedicating with acetaminophen, antihistamines, corticosteroids, and meperidine. Clinical experience has shown a patient intolerant of 1 amphotericin B lipid formulation may tolerate another well.
Nephrotoxicity is less common with all 3 lipid formulations of amphotericin B than with AmBD. In comparative studies the respective rates of nephrotoxicity (defined as twice baseline serum creatinine) are: ABLC ~25%, ABCD ~15%, L-AmB ~20%, and AmBD ~30% to 50%. Importantly, in open label studies, administration of lipid-based amphotericin B drugs are reported to stabilize or even improve renal function in patients with preexisting renal insufficiency. In addition, individuals receiving concurrent nephrotoxic agents (e.g., cyclosporine, aminoglycosides) are less likely to have renal dysfunction when receiving lipid formulations compared to AmBD. Hydration and sodium repletion have been utilized to reduce the risk of nephrotoxicity with AmBD but the utility of these measures is unclear with lipid formulations. Other adverse events associated with lipid-based products have included elevations in liver transaminases, alkaline phosphatase, and serum bilirubin levels. Liver function test abnormalities have been noted in 25% to 50% of patients treated with L-AmB but these findings are reversible without drug discontinuation in the majority of patients.
The costs of lipid formulations of amphotericin B are considerable. The daily treatment costs based on the 1999 average wholesale price for amphotericin B products are as follows: AmBD 1 mg/kg $25, ABLC 5 mg/kg $776, ABCD 4 mg/kg $480, and L-AmB 3 mg/kg $942. To our knowledge, no formal prospectively designed pharmacoeconomic (cost effectiveness) trials have been performed comparing lipid formulations to AmBD. Efforts by pharmaceutical manufacturers to reduce costs of their drugs are ongoing.
The 3 lipid-based formulations represent a significant advance in antifungal therapy. While superior clinical efficacy of these drugs over AmBD for any fungal disease has not been established, the lower toxicity profile of lipid-based agents is an attractive feature. All 3 drugs are less nephrotoxic than AmBD and 1 of the drugs (L-AmB) is clearly associated with fewer infusion-related adverse events than AmBD. The high cost of the lipid-based drugs is the major factor that limits their utilization and, as stated earlier, well designed pharmacoeconomic studies would be of significant value. FDA approved clinical indications (Table 3) are appropriate and justified by available data. In addition, many authorities recommend 1 lipid-based amphotericin B formulation as initial therapy for severe, life-threatening, invasive mold disease in immunocompromised patients. On the other hand, the first-line use of a lipid-based drug for empiric therapy in febrile neutropenic patients is more controversial. While some authorities advocate AmBD as initial therapy in this setting, others advocate a lipid-based drug. Cost effectiveness studies would be of considerable value in this patient population. The lipid-based drugs should not be used as initial polyene therapy for yeast infections (candidiasis and cryptococcosis) or endemic mycoses (blastomycosis, coccidioidomycosis, and histoplasmosis) unless amphotericin B is clearly indicated and the patient has preexisting renal dysfunction.
The oral triazoles, fluconazole and itraconazole, are active in vivo against clinically important yeast pathogens, including Cryptococcus neoformans, Candida albicans, and nonalbicans Candida species, as well as the common dimorphic fungi, including Coccidioides immitis, Histoplasma capsulatum, Blastomyces dermatitidis, Paracoccidioides brasiliensis, and Sporothrix schenckii. However, the triazoles, especially fluconazole, have significantly less activity against some of the nonalbicans Candida species, primarily Candida krusei and Candida glabrata. The widespread use of the triazoles has impacted the epidemiology and microbiology of candidal infections such that, over the past 10 years, there has been emergence of nonalbicans Candida species with significantly higher mean inhibitory concentrations to fluconazole compared to Candida albicans species. Itraconazole is unique in that it is the only azole compound with demonstrated efficacy for treating Aspergillus infections but should be used mainly in patients whose net state of immunosuppression has been stabilized or reduced. Several highly promising triazole antifungal drugs, including voriconazole and SCH 56592, are in various stages of development. Features that distinguish these new compounds from available azoles are their attractive pharmacokinetic profiles and broad spectrum antifungal activity, including enhanced in vitro and/or in vivo activity against Aspergillus species, fluconazole-resistant Candida species, and emerging mold pathogens including dematiaceous fungi. Phase II and III clinical trials are under way for both of these investigational drugs.
Craig J. Hoesley, MD
Instructor in Medicine
and
William E. Dismukes, MD
Professor and Vice-Chairman
Department of Medicine and
Director, Division of Infectious Diseases
University of Alabama School of Medicine at Birmingham
Birmingham, Alabama
