INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Amphotericin B (AmB) is an antifungal agent used as treatment for patients with established or suspected fungal infections and prophylaxis for immunosuppressed patients. Many children with malignant diseases who are admitted with fever receive empiric AmB therapy when the fever persists to prevent fungal superinfection and to control clinical or subclinical fungal infection. However, the side effects of AmB therapy are significant and include nephrotoxicity, electrolyte abnormalities (particularly hypokalemia), anemia, and infusion-related nausea, fever, and rigors (3, 7, 11, 12, 14, 31). Five clinical trials (9, 10, 18, 20, 24) with adults have compared the toxicity of AmB infused in lipid emulsion (AmB-IL) with that of the AmB infused in 5% dextrose (AmB-DX), which is conventionally used, and the results from three of these studies (9, 10, 20) have suggested that there is reduced nephrotoxicity and improved clinical tolerance when AmB is administered in lipid emulsion. Pharmacokinetic studies with adults (2, 15) have shown that the pharmacokinetics of AmB are different when it is administered in different infusion diluents. No studies have compared the toxicities and pharmacokinetics of AmB-IL and AmB-DX in children.

The aim of this study was to examine the toxicity and pharmacokinetics of AmB in children with malignant disease randomized to receive either conventional AmB-DX or AmB-IL (20% Intralipid).

We thank the patients and their families for taking part in the study, the nursing staff in the oncology unit for their care of the patients including the taking of samples for pharmacokinetic analysis, and the staff in the Department of Biochemistry for measuring the plasma creatinine, urea, and potassium concentrations.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Study design. This study was a prospective, single-center, randomized, open-label, controlled comparison of the clinical tolerance, toxicities, and pharmacokinetics of AmB-DX and AmB-IL. Restricted randomization was used in this study; the allocation arms were randomly assigned to consecutively numbered envelopes. The protocol was approved by the New Children's Hospital Ethics Committee. The parents of all children in the study gave informed consent.

Patients. All oncology patients who commenced AmB therapy, including those who had previous bone marrow transplants, were eligible for the study. Patients were allowed to be reentered into the study and rerandomized with subsequent courses of AmB. Patients were randomly assigned to one of the two AmB treatment groups (AmB-DX or AmB-IL).

Monitoring of side effects. Assessment of the extent of renal dysfunction was based on changes in plasma creatinine, urea, and potassium concentrations and in potassium requirement.

Formulation and administration of AmB. The AmB desoxycholate (Fungizone; Brystol-Myers-Squibb) dose, calculated as 1 mg/kg of body weight, was reconstituted with distilled water and was diluted with 5% dextrose solution (AmB-DX) or parenteral fat emulsion (20% Intralipid; Kabi Pharmacia) (AmB-IL). Doses of 40 mg of AmB or less were given in 100 ml of solution (i.e., maximum concentration of 0.4 mg/ml). Doses of over 40 mg (with a maximum of 50 mg) were given in 250 ml of solution. The AmB dose was administered as a single daily 2-h intravenous infusion. All patients were premedicated routinely with hydrocortisone (3 mg/kg). All patients had a double-lumen central line into the right atrium, so that one lumen could be used for drug administration and one could be used for sampling.

Blood samples for pharmacokinetic analysis. Heparinized blood samples (1 to 2 ml) for the measurement of the AmB concentration for the pharmacokinetic analysis were collected before the infusion and at 2 (infusion end), 3, 4, 6, 12, 18, and 24 h after the start of the infusion on the first day of therapy. For subsequent doses given once daily, blood samples for trough AmB level determinations were collected immediately prior to administration of the next AmB dose. Blood was centrifuged in a Beckman GS-6R centrifuge at 3,000 rpm for 10 min at 4°C, and plasma was stored at 40°C and analyzed within 1 week of collection. AmB has previously been shown to be stable in human plasma at 4 and 20°C for 24 days (13).

Assay of AmB. The AmB concentration in plasma was determined by a modified version of a previously reported method (13). To an aliquot of plasma (0.1 ml) in a microcentrifuge tube was added internal standard solution (20 µl of a 4-mg/liter solution of p-nitroaniline) and acetonitrile (0.25 ml). The tube was vortexed for 10 s and was allowed to stand for 5 min before centrifugation at 1,200 × g for 2 min, and a 50-µl sample of the acetonitrile extract was injected into the high-performance liquid chromatography (HPLC) system. The HPLC system consisted of an Altex model 100 pump and a Waters model 480 UV detector set at 405 nm with a Phenomenex Spherex (15 cm by 4.3 mm by 5 µm) C₁₈ column fitted with a Brownlee (1.5 cm by 3.2 mm by 7 µm) RP-18 precolumn. AmB eluted with a retention time of 9 min at room temperature when 1 ml (per min) of 5.6 mM sodium acetate buffer (pH 7.4) with 43.3% acetonitrile and 0.9% EDTA was used. The internal standard had a retention time of 4.5 min. The peak height ratio (ratio of the AmB peak height to the internal standard peak height) of the unknown sample was compared with the peak height ratios of spiked plasma standards containing 0, 0.5, 1.0, 1.5, and 2.0 µg of AmB per ml. The standard curve was demonstrated to be linear for between 0.1 and 6 µg of AmB per ml on several occasions. The between-day coefficient of variation of the assay was 15% for a concentration of 0.5 µg of AmB per ml (n = 83) and 11% for a concentration of 1.5 µg/ml (n = 83), and the limit of detection was 0.1 µg/ml.

Pharmacokinetic analysis. First-dose pharmacokinetic analysis was performed with AmB concentration-time data for 20 children who received AmB-DX and 15 children who received AmB-IL (see Table 2). Mean residence time (MRT) and the area under the plasma concentration-time curve (AUC) were estimated by the linear trapezoidal rule with extrapolation to infinity (23). Clearance (CL) was calculated by dividing the dose (in milligrams per kilogram) by the AUC extrapolated to infinity (AUC_0-). The steady-state volume of distribution (V_SS) was calculated as the product of CL and MRT. The terminal elimination rate constant (k_el) was determined from the slope of the terminal portion of the log concentration-time curve. There were at least three and up to five datum observations in the log-linear terminal phase. The slope of the terminal decline in the log-linear concentrations was determined by unweighted linear regression. The elimination half-life (t_1/2) (harmonic mean) was calculated by dividing the natural logarithm of 2 (0.693) by the mean k_el value. The initial volume of distribution (V) was calculated by dividing CL by k_el.

Statistical analysis. The analysis of the adverse effects of AmB was performed for two groups of patients. The first group included all patients who were enrolled in the study, according to the intent-to-treat principle (the whole cohort). The second group included all patients who received a full course of AmB for 6 days or more (the full-course group).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Patients. Table 1 compares the baseline characteristics of the groups of patients in the whole cohort and in the full-course group randomized to receive AmB-DX or AmB-IL. There was no significant difference in age, weight, height, surface area, gender, underlying disease, total parenteral nutrition administration, baseline creatinine concentration, days of daily AmB therapy, or AmB dose between the two treatment groups for both study populations (the whole cohort and the full-course group). Exposures to other potentially nephrotoxic agents were also similar. For the AmB-DX and AmB-IL treatment groups there were no significant differences in the numbers of courses after bone marrow transplantation or the numbers of courses in which cyclosporin or diuretics were used.

Clinical outcome. Of the whole cohort of 162 courses of AmB, positive fungal cultures were observed for 20 courses: there were 10 documented fungal infections in the AmB-DX treatment group (6 Candida spp., 1 Fusarium sp., 1 Aspergillus sp., 1 Scedosporium sp., and one unknown) and 10 documented fungal infections in the AmB-IL treatment group (all Candida spp.). One child died (AmB-DX arm) from a disseminated fungal infection (a 0.9% mortality rate). AmB appeared to be successful in all other instances in preventing or controlling fungal infection.

Pharmacokinetics. A semilogarithmic plot of the disposition of AmB in 20 children who received AmB-DX and in 15 children who received AmB-IL is shown in Fig. 1. Mean plasma AmB concentrations and 95% confidence intervals for the two groups of patients are shown for the first 24 h. Plasma AmB concentrations were generally lower for the AmB-IL treatment group than for the AmB-DX treatment group.

View larger version (21K): [in this window] [in a new window]   FIG. 1.   Semilogarithmic plot of mean AmB concentrations and 95% confidence intervals after administration of the first dose in dextrose or lipid emulsion in children with malignant disease.

1

1

1

1

1

1

1

1

1

1

Effects on markers of renal function. In Table 3, the mean plasma creatinine, urea, and potassium concentrations, as well as the potassium requirement, are shown for both the 1st and 7th days of therapy for the children who received the full course of AmB-DX and AmB-IL. There was wide interpatient variation in both plasma creatinine and urea levels on day 7, and most were within the normal reference interval of 30 to 75 µM for the plasma creatinine concentration in children ages 5 to 15 years. Only 2 (of 162) courses of AmB were associated with nephrotoxicity, which was defined previously (13) as a greater than 100% increase in the plasma creatinine level from the baseline to a level above normal (see Table 5). Despite the low incidence of AmB-induced nephrotoxicity, the results in Table 3 do provide evidence that renal function was affected in patients receiving both AmB-DX and AmB-IL. It was observed that there were significantly higher plasma creatinine and plasma urea concentrations and significantly higher potassium requirements on the 7th day of AmB therapy compared with those on the 1st day.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

This study compares the pharmacokinetics and side effects of AmB-DX with AmB-IL (20% Intralipid) in children with malignant disease. Pharmacokinetic studies were performed only after administration of the first dose of AmB, and model-independent pharmacokinetic analyses were used.

The pharmacokinetics of AmB-DX in children has been reported previously (4, 19, 25), and comparisons of previously published data with the results from this study are shown in Table 6. CL, volume of distribution (V), and elimination t_1/2 for AmB-DX for this study closely agreed with those of Benson and Nahata (4), who investigated AmB-DX infusion in a group of oncology patients with an age range similar to that of the patients in this study and who also used model-independent pharmacokinetic analysis. There were methodological differences between the various studies, with Starke et al. (25) and Koren et al. (19) using a one-compartment pharmacokinetic modeling approach and with Starke et al. (25) including 7 (of 10) patients under 1 year of age. The variation in AmB pharmacokinetics reported from these different studies may be due to these methodological differences.

The pharmacokinetics of AmB-IL differed considerably from those of AmB-DX. When AmB-IL was administered, the values for CL, V, and V_SS were 1.6-, 1.8-, and 1.8-fold higher, respectively, than those obtained when AmB-DX was administered. The C_max, C_SS,trough, and AUC_{0-24 h} values for AmB-IL were 0.68, 0.75, and 0.60 the values for AmB-DX, respectively. We found no significant difference between AmB-DX and AmB-IL C_24 h,troughs, but the numbers were smaller than those for the comparison of C_SS,troughs. The terminal t_1/2 and the MRT were also unaffected by the method of AmB administration. A previous study performed with adults (2) also showed that AmB-IL was associated with a 1.9-fold higher V_SS compared with that for AmB-DX, with the t_1/2 and MRT remaining constant.

It is possible that differences in AmB aggregate formation could contribute to the observed differences in the pharmacokinetics of AmB-IL and AmB-DX. It has been reported previously (26) that AmB does not mix well with lipid emulsion, with there being a greater degree of self-association of the AmB molecules in the lipid emulsion. Intralipid has a particle size of less than 1 µm (27), but Trissel (26) found that there were substantially more particles, many of them greater than 10 µm in diameter, in AmB-IL than in AmB-DX. Ranchere and colleagues (22) found that the particle size in an infusion containing AmB and Intralipid was up to fourfold greater than that in an infusion with dextrose and Intralipid. The possibility therefore exists that, as discussed previously (15, 30), AmB administered in lipid emulsion leads to an increased uptake of the drug by the organs of the reticuloendothelial system (liver, spleen, and lung) due to the larger particle size. This could account for the increased CL of the drug (especially uptake into the liver and subsequent hepatic elimination) and an increase in V_SS due to more extensive tissue uptake. In support of this, there appears to be increased uptake of AmB-IL into the lungs of mice compared with that of AmB-DX (17).

Another possible explanation for the increased V, V_SS, and CL associated with AmB-IL is that there may be a difference in the concentration of unbound AmB in AmB-IL compared with that in AmB-DX. It has previously been postulated (16) that the unbound concentration of AmB may differ widely between dose forms (dextrose versus lipid emulsion versus liposomal form), although this has not been investigated. The unbound fraction is more available for hepatic elimination and distribution into the tissues. AmB is highly bound in plasma (96.5% [21] and 91 to 95% [5]), mainly to plasma lipoproteins (6, 28, 29). The unbound AmB concentration is therefore likely to be very small, and a change could significantly affect the pharmacokinetics of AmB. Intralipid may compete with lipoproteins for the binding of AmB molecules, effectively increasing the fraction of AmB that is unbound. An increase in the unbound fraction could thus explain the observed increases in V, V_SS and CL that occurs when AmB is administered in lipid emulsion.

We observed that therapy with either AmB-DX or AmB-IL over 6 days was associated with significant increases in plasma urea and creatinine concentrations, increases in potassium requirement, and infusion-related fever and rigors. The magnitude of the change in creatinine was generally not very large, and nephrotoxicity was evident for only 2 (of 162) courses of AmB. Altered renal function is well known to be a consequence of AmB therapy (31), and the changes in creatinine and urea concentrations and potassium requirements observed after 6 days of AmB therapy appeared to be very useful markers of this side effect of AmB in children.

The low incidence of nephrotoxicity observed in this study may be due to a number of reasons. Patient age and underlying disease state may affect the pharmacokinetics of AmB in such a way that fewer side effects are manifested. The fact that patients were premedicated with hydrocortisone and were not sodium depleted at the initiation of AmB therapy may also have contributed to the low incidence of side effects.

The changes in potassium requirement and plasma urea or plasma creatinine concentration after 6 days of AmB therapy were not significantly different for AmB-DX and AmB-IL, indicating that there was no difference in the way that they affected renal function, despite the lower AUC for AmB-IL.

In an intent-to-treat analysis there was also no detectable difference between AmB-DX and AmB-IL in the numbers of children experiencing nephrotoxicity or infusion-related fever and rigors (although not all patients had a full follow-up with respect to this end point). Previous studies with adults have shown that AmB-IL is better tolerated than AmB-DX, with fewer patients experiencing adverse reactions (9, 10, 20), but we found no such difference in children. Joly and colleagues (18) also found that AmB-IL did not have substantially reduced renal toxicity compared with AmB-DX in AIDS patients, although Intralipid infusion did reduce the infusion-related toxicity of AmB without altering the antifungal efficacy. In addition, a recent investigation comparing the safety of AmB-DX and AmB-IL in adult neutropenic patients found that administration of AmB-IL provided no benefits and was associated with increased occurrences of acute pulmonary side effects (24).

This study, in agreement with previous clinical trials (8, 10, 18), showed that AmB-IL and AmB-DX are equally effective in preventing or controlling fungal infections. However, only 20 episodes of cultures positive for fungal infection were identified, and there were insufficient numbers for a good comparison of efficacy. It is possible that the efficacy of AmB-IL could be compromised by the observed lower plasma AmB concentrations because it has previously been suggested that maintenance of serum AmB concentrations above the MIC for the invading organism is a minimum requirement for successful therapy (1).

The low incidence of AmB-induced nephrotoxicity in this study was a limitation in the detection of differences in toxicity between AmB-DX and AmB-IL. However, this study has demonstrated that there is no advantage to administering AmB in lipid emulsion in children with malignant diseases, a patient group that tolerates AmB therapy well.

There are now reports that suggest that the lipid preparation of AmB is not suitable for routine use: the product is inconsistent and undesirable (22, 26, 30). Until it becomes clear why administration of AmB in lipid emulsion has such a marked effect on the pharmacokinetic parameters, including AUC, C_SS,trough, CL, and V_SS, and yet does not alter the pharmacodynamic parameters, including the changes in the plasma creatinine and plasma urea concentrations and the potassium requirement, the use of lipid emulsion for the administration of AmB must remain suspect. Further work comparing the unbound concentrations of AmB, the tissue distribution of AmB, and the distribution of AmB to the lipoprotein fractions of blood for AmB-DX and AmB-IL is needed to explain these effects, but in view of the possible safety concerns with AmB-IL, these may not be warranted.

In conclusion, this investigation found no significant advantage in safety or tolerance by administering AmB in lipid emulsion compared to administering it in dextrose in the treatment and prevention of fungal infection in children with malignant disease. In view of this finding and the reported incompatibility of AmB with lipid emulsion (26), we recommend the use of only conventional AmB-DX.