The incidence of fungal sepsis in the neonatal intensive care unit (NICU) has increased recently,1,2 and fungal sepsis is the third most common cause of late onset sepsis in the NICU.3 Systemic candidiasis is the most common and fatal cause of fungal sepsis in newborns.4,5 Moreover, 2-4% of neonates with systemic candidiasis are very low birth weight infants (VLBWI), weighing less than 1500g, and have a high mortality rate.6,7 Major risk factors for systemic candidiasis include intravenous catheter use, prolonged antibiotic therapy (particularly third-generation cephalosporin usage), mechanical ventilation, prolonged total parenteral nutrition, and prolonged hospitalization. Despite improvements in neonatal intensive care, the mortality rate in VLBWI with systemic candidiasis receiving antifungal agents is high.1 Currently amphotericin B is considered to be the treatment of choice for systemic candidiasis but its use is limited due to adverse effects such as renal toxicity, hepatotoxicity, electrolyte imbalances, and bone marrow suppression.8,9 Preterm infants born before 34 weeks of gestation have a significantly lower glomerular filtration rate (GFR) and renal tubular function than full-term infants, which persists for up to 3-5 weeks post-natally.10 In addition, many risk factors including the maternal use of non-steroidal anti-inflammatory drugs during pregnancy, respiratory distress syndrome (RDS), a low Apgar score and ibuprofen treatment are associated with impaired renal function in preterm infants.11 The renal function of these preterm infants treated with amphotericin B is even further deteriorated as a result of renal tubular dysfunction and decreased GFR due to amphotericin B induced vasoconstriction. This deterioration in renal function may limit amphotericin B use at the proper and maximum dosage, which can cause treatment failure and a high mortality rate.

New lipid formulated amphotericin B preparations have been reported to have reduced renal toxicity compared to conventional amphotericin B while maintaining the same effectiveness, and may be a more useful alternative in patients with impaired renal function. AmBisome^® is one of the lipid formulations of amphotericin B that has received considerable attention for its use in adult patients. This drug has received Food and Drug Administration (FDA) approval for the treatment of patients with 1) cryptococcal meningitis-HIV infection; 2) mycosis as an empiric therapy for a presumed fungal infection in patients with febrile neutropenia; 3) systemic mycosis due to aspergillus, candida, and cryptococcus in patients refractory to amphotericin B deoxycholate or where renal impairment or unacceptable toxicity precludes the use of amphotericin B deoxycholate; and 4) visceral leishmaniasis.1,12 Thus far, however, there is limited experience with the use of AmBisome^® in preterm infants with systemic candidiasis, and the few reports have been limited to uncontrolled, retrospective analyses and case reports. Therefore, amphotericin B continues to be the treatment of choice in preterm infants with systemic candidiasis. This study compared the safety and efficacy of AmBisome^® in the treatment of VLBWI with systemic candidiasis to that of conventional amphotericin B.

This study was a prospective historical control multi-center study and was reviewed, approved, and funded by the institutional review board of the IN-SUNG Foundation of Medical Research (C-A4-824-1). AmBisome^® is not covered by current medical insurance, and only patients with informed parental consent were enrolled in this study. Twenty-six VLBWI were treated with AmBisome^® (AmBisome group) for systemic candidiasis in the NICU from October 2003 to July 2006 at various medical centers. These medical centers included the Samsung Medical Center, Sungkyunkwan University School of Medicine (n = 18), Bucheon Hospital, Soonchunhyang University, College of Medicine (n = 3), Ilsan Paik Hospital, Inje University School of Medicine (n = 2), Ansan Hospital, Korea University School of Medicine (n = 2), and the Kangnam Cha Hospital, Pochon Cha University, College of Medicine (n = 1). VLBWI treated with amphotericin B, and who were from the same hospitals, were assigned as the historical control (n = 20, Amphotericin group) and were compared with the AmBisome group. Table 1 shows details regarding enrolled hospitals.

Systemic candidiasis was defined only when the blood culture was positive for Candida spp. with the clinical signs of a systemic infection. A blood culture positive for Candida spp. was mandatory but other cultures positive for Candida spp. such as urine, cerebrospinal fluid and tracheal aspirate were not required. Patients with clinical candidiasis enrolled in this study were defined as those having negative blood cultures with a suspicion of sepsis due to clinical and/or laboratory findings who did not respond to antibiotics for more than seven days. Blood (1/2mL) was extracted from patients using a sterile technique. The blood was then inoculated into aerobic and anaerobic blood culture media (BacT/ALERT^®, BioMerieux, Durham, NC, USA) and processed according to standard microbiologic techniques. Results of the blood cultures were initially reported one week after inoculation, and a final report was issued three weeks following inoculation.

Liposomal amphotericin B (AmBisome^®, Gilead Sciences International Ltd., Cambridge, UK) was administered with a beginning dosage of 1-3mg/kg/day which incrementally increased daily by 1-3mg/kg/day until maximum dosage at 5mg/kg/day. By this means, the target dosage was reached by the fifth day of administration with a final concentration of 1mg/mL diluted in 5% dextrose water over a continuous two hours infusion. Amphotericin B (Fungizone^®, Bristol-Myers Squibb, France) was administered with a daily dosage from 0.5-1.0mg/kg/day with a final concentration of 1mg/mL with 5% dextrose water over a continuous six hours infusion. The treatment was finished if the patients were considered to be clinically cured and received therapy until a negative blood culture from positive blood culture was observed, or if the patients with clinical candidiasis showed improved clinical and/or laboratory findings which means they responded to the antifungal therapy.

The efficacy of the two drugs was evaluated based on fungal eradication and mortality rates. Fungal eradication (resolution) means that a negative culture after the initiation of therapy in whom Candida spp. had previously been cultured from blood and urine, cerebrospinal fluid or transtracheal aspirate before antifungal therapy. The fungal eradication rate, fungal eradication time and the duration of anti-fungal therapy in the two groups were compared.

The incidence of renal toxicity, hepatotoxicity, hypokalemia, thrombocytopenia and fever in the two groups were evaluated to determine side effects and safety. The concentrations of creatinine, electrolytes, transaminase levels and urine output were determined before the onset of treatment, 24 h following the last dosage and at peak levels following treatment. The tests were repeated during treatment when deemed necessary. A renal side effect was defined as a 50% increase in the serum creatinine levels from the baseline level (renal toxicity),13 or oliguria, with urine output less than 1mL/kg/h. Liver side effects (hepatotoxicity) were defined as a 100% increase in the serum transaminase, particularly alanine aminotransferase (ALT) levels from the baseline level.13 Hypokalemia was diagnosed when the serum potassium concentration was < 3mmol/L, and thrombocytopenia was diagnosed when the serum platelet count was < 100 × 10⁹/L. A fever was defined as a 6 a.m. oral temperature of > 37.2℃ (> 98.9˚F) or a 4-6 p.m. oral temperature of > 37.7℃ (> 99.9˚F).14 Axillary temperatures were generally 0.4℃ (0.7˚F) lower than the oral readings, so a fever was defined as an axillary temperature of > 37.4℃ (≥ 37.5℃) at any time of the day.

Several perinatal and postnatal variables were evaluated to determine factors influencing the treatment outcome. These variables included: gestational age, birth weight, central line insertion, antibiotics use, duration of total parenteral nutrition, length of time on mechanical ventilation, length of hospital stay, gender, RDS, patent ductus arteriosus (PDA) confirmed with echocardiography, high grade intraventricular hemorrhage (IVH, ≥ Gr.III), periventricular leukomalacia (PVL), and bronchopulmonary dysplasia (BPD), oxygen dependency at a corrected age of 36 weeks.

The results are presented as the mean ± standard deviation. Statistical analysis was performed using a chi-square test or Fisher's exact test for dichotomous outcome data, and a t-test or Mann-Whitney U test for continuous data. Differences between AmBisome group and Amphotericin group were analyzed. A p value < 0.05 indicated a significant difference.

These results suggest that AmBisome^® is an effective and safe antifungal agent in the treatment of systemic candidiasis in VLBWI. The mean duration of AmBisome^® therapy was shorter than amphotericin B therapy, but fungal eradication rates, eradication time, and mortality rates were similar. AmBisome^® was administered at a higher daily maximum dose and cumulative dose than amphotericin B, but renal toxicity and hepatotoxicity were significantly lower in the AmBisome group.

Amphotericin B binds to ergosterol, a sterol component unique to fungal cell membranes, altering the cell permeability and causing the leakage of cytoplasmic content (electrolyte loss) that ultimately produces fungal death. Unfortunately, when amphotericin B is infused into circulation, it also binds to cholesterol, a sterol component of the mammalian cell membrane, and creates a similar disruption to the cell membrane. Again, the normal cytoplasmic contents leak, causing cell death that results in serious side effects such as renal toxicity.15 Although there was no detectable hepatic failure or renal failure caused directly by the drug itself, there was significant detectable hepatic toxicity or renal toxicity within the study. New lipid formulated amphotericin B preparations have been developed to reduce renal toxicity levels of amphotericin B and AmBisome^® is the only commercially available lipid formulation of amphotericin B containing liposomal structures. AmBisome^®, or liposomal amphotericin B, is a very small unilamellar vesicle (< 100nm in size) with amphotericin B intercalated within the phospholipid bi-layer of the liposome. AmBisome^® has a preferential affinity for fungal ergosterol cells over mammalian cholesterol cells. Therefore, once AmBisome^® binds to the fungal cell wall, the fungal cell releases phospholipase, which then digests the liposome. This causes amphotericin B, which is intercalated within the liposome, to leave the liposome and disrupt the fungal cell membrane by binding to it. Ultimately, this results in selective fungal cell death with fewer systemic side effects.16,17

Lipid formulated amphotericin B has several advantages over conventional amphotericin B, including an increased daily dosage of the parent drug (up to 10-fold), high tissue concentrations in the primary reticuloendothelial organs (lungs, liver, and spleen), a decrease in infusion-related side effects (especially liposomal amphotericin B), and a marked decrease in renal toxicity. AmBisome^® concentration was found to be highest in the liver, spleen, kidneys and lungs, and the extent of the tissue distribution of AmBisome^® may be an important determinant of the treatment outcome.18

The pathophysiology of renal toxicity for amphotericin B, other than non-selectively binding to mammalian cholesterol cells, involves the vasoconstriction of and direct interaction with the renal epithelial cell membranes, resulting in a decrease in GFR and tubular dysfunction. Amphotericin B forms pores in cell membranes causing tubular dysfunction, and is responsible for severe vasoconstriction that decreases renal blood flow and GFR, ultimately causing ischemic injury. Together, these two mechanisms induce acute renal dysfunction. Several papers report the rate of acute renal failure for patients on amphotericin B to be between 49% and 65%.9 Amphotericin B induces the following changes: hypokalemia, hypomagnesemia, renal tubular acidosis, and increases in serum creatinine preceded by tubular dysfunction. In addition, amphotericin B is responsible for prolonged hospitalization, as well as increased costs and mortality due to hemodialysis caused by amphotericin B induced acute renal failure. Therefore, some studies regard this format of renal toxicity not as a simple, transient episode but a severe disease that should be prevented if at all possible.19 Scarcella et al.20 concluded, however, that amphotericin B is well-tolerated in preterm infants. Recently, many specialists have grown to regard the most common side effect of amphotericin B as mild renal toxicity that is resolved after the cessation of therapy.21 In our study, renal function was evaluated based on urine output and serum creatinine levels. Urine output and episodes of oliguria following treatment were similar between the study groups. Serum creatinine levels were similar for both groups, not only before treatment but also at the peak level following treatment. Additionally, renal toxicity, defined as a 50% increase in the serum creatinine level, was lower in the AmBisome group following treatment.

The ALT level before the onset of treatment was similar for both groups, but the peak level following treatment was higher in the Amphotericin group. The incidence of hepatotoxicity, defined as a 100% increase in the serum ALT level, was significantly lower in the AmBisome group. Drug-induced hepatotoxicity usually presents symptoms of acute hepatocellular injury, cholestasis, or both. Generally, in hepatocellular disorders, ALT concentration is elevated, and in cholestatic injury, bilirubin and alkaline phosphatase concentrations may or may not be elevated. A case report describes the hepatotoxicity with amphotericin B in an adult patient.22 A 53-year-old woman with a non-small cell lung cancer received amphotericin B, and on the next day, her bilirubin, aspartate aminotransferase (AST), and ALT levels had increased markedly. These levels were normalized after discontinuing amphotericin B and AmBisome^® was then administered. Following AmBisome administration, bilirubin levels increased mildly, but the AST and ALT were unaffected. The bilirubin level normalized following discontinuation of the drug. However, the mechanism of hepatotoxicity induced by amphotericin B was not determined in this study. Mohan and Bush23 (2002) reported the first case of amphotericin B induced hepatotoxicity in children. The patient was a 9-year-old girl with cystic fibrosis who had developed fulminant renal and hepatic dysfunction after a short course of intravenous amphotericin B for a suspected aspergillus infection, even though she did not have clinical evidence of invasive aspergillosis. The patient's maximum AST and ALT levels were up to 1112 U/L and 364U/L, respectively. Her AST and ALT improved gradually after discontinuing the amphotericin B and normalized after a seven-day period. Juster-Reicher et al.1 reported a transient increase in the serum bilirubin and hepatic transaminases levels in one infant after administering AmBisome^® in 24 VLBWI, all weighing less than 1500g. They could not explain the mechanism of AmBisome^®-induced hepatotoxicity. Some researchers23 have reported that the non-selective disruption of mammalian cells due to cholesterol binding is the mechanism for the toxic effects of amphotericin B, but the precise mechanism of hepatotoxicity is unknown despite several case reports of amphotericin B and AmBisome^®-associated hepatotoxicity.

In one study, a comparison of amphotericin B therapy (n = 34) and AmBisome^® therapy (n = 6) in VLBWI weighing less than 1500g24 concluded that the fungal eradication time was five days for the AmBisome group, and eight days for the Amphotericin group. In addition, the fungal eradication rate was 83% in the AmBisome group, and 68% in the Amphotericin group. Mortality due to the fungal infection was 16.7% and 14.7%, respectively. In this study, the number of VLBWI assigned to the AmBisome group (n = 26) was higher than that reported by Linder et al. (n = 6),24 the fungal eradication rate was higher (84% in AmBisome group, 89% in Amphotericin group), and the mortality was lower (12% in AmBisome group, 10% in Amphotericin group). Scarcella et al.20 administered liposomal amphotericin B to 40 preterm neonates with a severe fungal infection. The dose ranged from 1mg/kg/day to a maximum of 5mg/day based on the clinical response. Fungal eradication was observed in 70% (28/40) of the patients.

In conclusion, AmBisome^® is an effective and safe antifungal agent for the treatment of systemic candidiasis in VLBWI weighing less than 1500g.