INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Amphotericin B (AmB) has been available for more than 30 years and remains the agent of choice for the treatment of many severe, life-threatening fungal infections. However, use of this drug is often associated with adverse reactions, which may be classified as either acute or delayed. Acute (infusion-related) toxicity has been observed in up to 79% of patients receiving AmB and may include fever, chills, headache, nausea, vomiting, hypotension, bronchospasm, or allergic reactions (8, 15, 21, 25). Delayed toxicity also occurs commonly and may include anemia, nephrotoxicity, hypokalemia, or hypomagnesemia (21).

Recent studies have suggested that AmB incorporated into liposomes or mixed with Intralipid (IL) may have a decreased risk of both acute and delayed toxicity (12, 14, 20, 22, 29, 35). Interest in studying the efficacy and safety of liposomal AmB and of AmB-IL has arisen because of the observation that these compounds may deliver AmB to its target site in the fungal cell membrane more selectively (12, 35). IL is an oil-water system that is based on vegetable oils (10 to 20%, vol/vol) stabilized by phosphatides (1 to 2%, wt/vol) (5). Lipophilic substances can be solubilized in the oil phase, and the lipid emulsion can also stabilize drugs which are not stable in an aqueous environment (5). IL has been used as a carrier for lipophilic drugs in an attempt to improve drug stability and decrease the occurrence of adverse effects. In several small trials using AmB-IL for the treatment of serious fungal infections, a reduction in acute and delayed toxicity has been reported; compared to AmB administered conventionally in 5% dextrose in water (D5W) (1, 2, 4, 5, 10, 23, 28, 30).

Although AmB has been found to be stable in D5W for up to 5 days (18), only limited data are available regarding the stability and in vitro antifungal activity of AmB in IL. While a number of investigators have reported that AmB does not degrade while stored in IL (19, 26), others have reported the presence of particles (27, 32), precipitation of AmB (9, 27, 32), and reduced delivery of the dose (7). Knowledge of stability and compatibility and assurance that the mixture retains antifungal activity are essential prior to administration of AmB-IL to patients. Therefore, this study was conducted to determine the stability and compatibility of AmB-IL, as well as to compare the in vitro antifungal activity of AmB-IL to that of AmB alone against reference strains of Candida species.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Assay validation. AmB (Fungizone; Bristol-Myers Squibb Pharmaceutical Group, Montreal, Quebec, Canada) was intentionally degraded to ensure that the high-pressure liquid chromatography assay developed was capable of separating the drug from its degradation products, according to stability indicating procedures (31, 33). AmB (100 mg) was dissolved in 50 ml of sterile water to make a 2-mg/ml stock solution (pH 7.5). Ten-milliliter aliquots of the stock solution were placed in vials, and the pH was adjusted to 1.5, 7.5, or 12.5 by using 1 M hydrochloric acid or 1 M sodium hydroxide. Each solution was incubated in a water bath at 81°C for 5 h. Samples of each solution were drawn for chromatographic analysis just prior to incubation and at six other times during the period of incubation. Chromatograms were inspected for the appearance of additional peaks (degradation products of AmB), and the AmB peaks were compared between samples for changes in concentration, shape, or retention time.

Chromatography. The chromatographic conditions capable of separating AmB from its degradation products utilized an isocratic mobile phase consisting of 48% acetonitrile and 52% 0.05 M phosphoric acid. The solvent was pumped through a reverse-phase C₁₈ (4.6 by 100 mm) 3-µm analytical column (Partisil 5 ODS-3; Whatman Inc., Fairfield, N.J.) at a flow rate of 1 ml/min. A guard column (Aquapore RP-18 ODS; Applied Biosystems, Foster City, Calif.) was also used. Under these conditions, AmB eluted at 2.9 min. AmB was detected by using a variable-wavelength UV detector (model 759A; Applied Biosystems) at 220 nm. Chromatograms were recorded on a system using PC1000 software (Thermo Separation Products, Chicago, Ill.).

Sample preparation and assay validation. IL (Pharmacia Inc., Mississauga, Ontario, Canada) is a liquid emulsion containing either 10 g (10%) or 20 g (20%) of purified soybean oil, 1.2 g of purified egg phospholipids, and 2.2 g of anhydrous glycerol with sterile water. Tetrahydrofuran (1 ml) was added to 0.5-ml aliquots of IL containing AmB concentrations ranging from 0.05 to 1.5 mg/ml. This mixture was vortexed for 30 s. A sample of the tetrahydrofuran layer was injected into the high-pressure liquid chromatography column (Partisil 5 ODS-3; Whatman Inc.). Absolute recovery of this extraction was evaluated against six AmB standards in water compared to six AmB samples extracted from IL. The recovery was also evaluated against blanks to ensure that additional peaks did not interfere with AmB quantification.

Compatibility of AmB with IL. A compatibility study was done to evaluate the effects of the soybean oil concentration and the AmB concentration (32). AmB (50 mg) was reconstituted with 10 ml of sterile water, resulting in a 5-mg/ml solution. Thirty-two mixtures with various proportions of IL in combination with the reconstituted AmB in D5W were prepared. All mixtures had a final AmB concentration of either 0.6 or 1.2 mg/ml. A series of control samples without AmB were also prepared. Following mixing, the samples were allowed to stand for 1 h and then centrifuged for 30 min. The clear lower aqueous phase was visually inspected for the presence of a yellow precipitate of AmB indicating incompatibility (32).

Stability of AmB in IL. The stability of AmB in IL was tested over 21 days. Sixteen different AmB-IL solutions previously determined to be compatible were prepared for analysis to evaluate the effects of the AmB concentration (1.0 and 0.1 mg/liter), the IL concentration (10 and 20%), the temperature (4 and 23°C), and shaking. A stock solution of AmB in sterile water (5 mg/ml) was prepared. To prepare 1.0-mg/ml solutions of AmB, 36 ml of the AmB stock solution was added to 90 ml of D5W and 54 ml of 10 or 20% IL. To prepare 0.1-mg/ml solutions of AmB, 3.6 ml of the AmB stock solution was added to 122.4 ml of D5W and 54 ml of 10 or 20% IL. After recording the initial physical appearance, the 1.0- and 0.1-mg/ml solutions were each divided into 12 aliquots. Half of the solutions were stored at 4°C, and half were stored at 23°C. In addition, half of the solutions were shaken prior to sampling and half were not shaken. The solutions were stored for 21 days, and on each study day (days 0, 2, 5, 7, 9, 13, 15, 19, and 21) the solutions were physically inspected and the pH and AmB concentration were determined by obtaining samples from the surface of the mixture. Also on each of these days, fresh AmB-IL standards were chromatographed to construct a standard curve. The peak areas were subjected to least squares linear regression, and the actual AmB concentrations (from the average of two replicates from each sample) were interpolated from these curves and recorded. Concentrations were recorded to the nearest 0.001 mg/ml.

In vitro antifungal activity of AmB-IL. Analytical-grade AmB (Bristol-Myers Squibb) was prepared as a stock solution at a concentration of 2,560 µg/ml in dimethyl sulfoxide (Sigma Chemical Co., St. Louis, Mo.). RPMI 1640 medium (Sigma Chemical Co.) without bicarbonate was prepared by following National Committee for Clinical Laboratory Standards guidelines (24). The medium was buffered to pH 7.0 with 0.165 M morpholinepropanesulfonic acid (Sigma Chemical Co.). Serial dilutions of AmB in RPMI 1640 medium with IL containing 10 or 20% soybean oil or with no IL were made for macrobroth dilution susceptibility testing (24).

Statistical analysis. Reproducibility was assessed by calculating the CV. Mean results from the same test done on different days were compared by least-squares linear regression to determine if an association existed between the observed result and time. AmB concentrations were considered to be within acceptable limits if the concentration on any day of analysis was at least 90% of the initial concentration. The components of a solution were judged to be physically compatible if there was no visual change in the color of the mixture and no precipitate or separation of phases or particulate formation was observed.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Assay validation. At 81°C, a 2-mg/ml solution of AmB at pH 7.5 appeared to degrade very slowly, retaining more than 91% of the initial concentration in the 5-h study period. However, at pHs 1.5 and 12.5, the AmB solutions degraded to less than 10% of the initial concentrations within 30 min (Fig. 1). Degradation products did not interfere with AmB quantification. The standard curves of AmB-IL were linear for AmB concentrations ranging from 0.05 to 1.5 mg/ml (r > 0.98). The mean relative error (CV) observed in the analysis of duplicate AmB-IL samples between days ranged from 0.8 to 1.9%. The mean errors of duplicate analyses done within a day ranged from 0.1 to 3.2%. The lower limit of detection of AmB in IL was 0.05 mg/ml, as this was the lowest prepared standard concentration. At this concentration, the mean relative error was 1.8%.

View larger version (12K): [in this window] [in a new window]   FIG. 1.   Degradation curves of AmB in sterile water (2 mg/ml) at 81°C and pHs 1.5, 7.5, and 12.5.

Compatibility. The results of testing for the compatibility of AmB and IL are summarized in Table 1 and were found to be dependent on the concentration of soybean oil in the IL and the concentration of AmB or IL in the final mixture. A precipitate was observed following centrifugation in most mixtures prepared with IL with 20% soybean oil content and having a final AmB concentration of 1.2 mg/ml. When the concentration of AmB in mixtures prepared with IL containing 20% soybean oil was reduced to 0.6 mg/ml, a precipitate was observed in all mixtures with more than 30 ml of IL/100 ml of the total volume. With mixtures prepared with IL containing 10% soybean oil and having a final AmB concentration of 1.2 mg/ml, a precipitate was observed in all mixtures containing more than 30 ml of IL/100 ml of the total volume; when the concentration of AmB was 0.6 mg/ml, a precipitate occurred in mixtures with more than 40 ml of IL/100 ml of the total volume (Table 1).

Stability. Within 24 h of preparation, AmB separated from IL, forming a distinct lower yellow layer. The AmB could be readily resuspended within the IL by shaking, and the concentration of AmB in all samples which were shaken before sampling retained greater than 90% of the initial concentration for 21 days when stored at either 4 or 23°C (Fig. 2A and B). Conversely, in samples which were stored at room temperature and not shaken, the AmB concentration declined during the 21-day study period (Fig. 2A).

View larger version (17K): [in this window] [in a new window]   FIG. 2.   Effects of temperature, shaking, and IL and AmB concentrations on the stability of AmB suspended in IL. (A) Stability of AmB at 0.1 and 1.0 mg/ml and 23°C. (B) Stability of AmB at 0.1 and 1.0 mg/ml and 4°C.

In vitro antifungal activity of AmB-IL. The results of susceptibility testing with various concentrations of AmB in IL are summarized in Table 2. The results indicate that the presence of IL did not reduce the in vitro activity of AmB against the yeast strains tested. Repeat testing showed that these results were reproducible (data not shown).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

AmB has previously been shown to be stable in D5W for at least 5 days (18). The results of this study indicate that AmB in IL is stable for up to 21 days when stored at either room temperature or 4°C. Previous reports have also indicated that mixed AmB and IL are physically compatible and stable for over 24 h at AmB concentrations of up to 2 mg/ml (19, 26, 27). Although some separation has been reported to occur within AmB-IL mixtures over 24 h, this has not been considered to be an indication of incompatibility because gentle shaking produced a homogeneous mixture containing the expected concentrations of AmB. However, precipitation of AmB in AmB-IL mixtures has also been reported to occur (7, 9, 27, 32, 34). The observation by Trissel (32) that the precipitate consists of particles greater than 10 µm in size suggests that even without centrifugation, certain AmB-IL mixtures might be unsafe for human administration. It has also been suggested that the formation of these aggregates might result in decreased antifungal activity (16). Therefore, we evaluated the concentration-dependent compatibility of AmB-IL mixtures subjected to centrifugation. The design of the current stability study was based on the results of the compatibility evaluation, which demonstrated the physical compatibility of AmB mixtures containing less than 30 ml of IL/100 ml of the total volume. By increasing the volume of D5W and thereby decreasing the percentage of lipid in the total volume to less than 30%, the formation of a precipitate after centrifugation could be avoided (Table 1). Only under these conditions would infusion of AmB-IL be considered safe. However, our analysis did not include a determination of particle size. Verification of particle size in these mixtures should be done before a definitive recommendation regarding product safety can be made.

Mixing AmB in IL did not appear to adversely affect the in vitro activity of AmB against the small number of Candida strains evaluated in this study. In fact, our results were similar to those of Chavanet et al. (3), in that there was a slight reduction in the MICs when AmB was suspended in IL rather than D5W. These results need to be validated by evaluating a larger sample of yeasts.

AmB-IL formulations were similar in in vivo activity but less toxic than conventional AmB in a murine model of candidiasis (17) and cryptococcosis (13). Preliminary clinical studies (1, 2, 4, 23, 28, 30) with AmB-IL and a small number of patients (reviewed by Sievers et al. [29]) have also indicated equivalent efficacy but reduced toxicity compared to standard administration of AmB for the treatment of a variety of systemic fungal infections. Reduced toxicity was often observed despite the use of higher doses of AmB in IL (2, 3, 11, 23). It has been suggested that the observed reduced toxicity may be due to reduced delivery of AmB from AmB-IL mixtures (7). Chavanet et al. (4) and Heinemann et al. (10) monitored AmB concentrations in serum after administration of AmB in lipid emulsions and found lower peak concentrations in serum and more rapid clearance of AmB-IL than AmB alone. While these observations (4, 10) are consistent with reduced delivery of AmB, they are similar to findings reported with some liposomal AmB formulations without precipitate formation or reduced delivery of the drug (6). There is a need for clinical trials with larger numbers of patients with a variety of fungal infections to better document the efficacy and safety of AmB-IL. However, based on the results of this study and that by Trissel (32), AmB should not be combined with IL in concentrations or proportions (as identified in Table 1) capable of producing a precipitate.