Deoxycholate Amphotericin B and Nephrotoxicity in the Pediatric Setting

Abstract

Since the introduction of amphotericin B as an antifungal agent, the morbidity and mortality of pediatric patients with mycotic infections have increased, primarily because of the increased immunocompromised patients. Despite the fact that deoxycholate amphotericin B was once the primary drug used for mycotic infections, its administration to children older than neonates is currently controversial because of its nephrotoxic effects. Three lipid-associated formulations have been developed and have reportedly shown similar efficacy and fewer nephrotoxic effects in adults than conventional amphotericin B, but the conclusions from comparative studies in children evaluating the nephrotoxicity risks of the 4 agents are controversial. Nevertheless, guidelines favor liposomal or lipid complex amphotericin B when polyene antifungal therapy is recommended in this age group. However, high acquisition costs often preclude their prescription in economically poor regions. Thus, physicians must consider all of these factors when determining the most cost-effective polyene antifungal treatment for their pediatric patients. This is particularly pertinent in developing countries where resources are scarce. Adjuvant sodium supplementation has been reported to be effective in protecting kidney function in extremely low birth weight infants prescribed deoxycholate amphotericin B. Further pharmacokinetic and pharmacodynamic studies of the drug in children could also provide information for rational dosing regimens designed to decrease nephrotoxicity. Conventional amphotericin B, with appropriate kidney protective measures, still plays a role in the treatment of empiric invasive mycotic infections in most pediatric patients. Liposomal and lipid complex amphotericin B should be reserved for those receiving long-term nephrotoxic agents or with altered renal function or disease. Antifungal susceptibility, renal compromise and the clinical status of the patient should determine treatment for culture-proven infections. Under the current cost limitations, undertaking and evaluating low-cost, kidney-sparing, deoxycholate amphotericin B treatments for children should be a primary concern.

Since the introduction of amphotericin B as an antifungal agent in the late 1950s, the morbidity and mortality of pediatric patients with mycotic infections have increased.^1–5 This increase is because of the increasing number of immunosuppressed patients resulting from a rise in oncologic treatments, transplantations, AIDS incidence, antibiotic use and very low birth weight neonates.^3–6 Additionally, the spectrum of fungal species has expanded.⁴

Although numerous antifungal agents have been approved for use, deoxycholate amphotericin B has been considered the “standard” therapy for most invasive fungal infections and the standard of comparison for newer antifungal agents.^7,8 However, deoxycholate amphotericin B has several drawbacks, such as infusion-related events (ie, chills, fever, headache, nausea and vomiting),^3,4,9 and most significantly, dose limiting nephrotoxicity.^3–5,8,10 These findings have led authors to reassess its preeminence in antimycotic therapy and comparative studies.¹¹

Amphotericin B-induced nephrotoxicity, although reversible, has been associated with increased morbidity, mortality and hemodialysis as well as protracted hospital stays and higher costs.^12–15 Furthermore, nephrotoxicity early in life can result in major adverse effects on the long-term well-being of children.⁵

To circumvent its adverse effects, amphotericin B has been combined with parenteral fat emulsion, but this combination has not been proven to be more effective or less nephrotoxic in children.¹⁶ Moreover, although the efficacy of this combination is comparable with that of conventional amphotericin B against certain strains of Candida sp in vitro, its use is not recommended because of the enlarged lipid particles that are formed and its variable stability in solution.^17,18

Three commercially produced amphotericin B lipid-associated formulations were released in the 1990s (ie, amphotericin B lipid complex, amphotericin B colloidal dispersion and liposomal amphotericin B) and showed efficacy similar to and nephrotoxicity lower than that of conventional amphotericin B.^1,8,14,19,20 The primary disadvantage of these formulations is their expense; the financial cost for daily treatment is from 25 to 230 times that of conventional amphotericin B treatment.^1,19,21–23

Although deoxycholate amphotericin B-associated nephrotoxicity rates in children and adults have recently been found to be comparable,²³ compromised kidney function has generally been reported to be less severe in children.^6,7,9,24,25 Nevertheless, the observed nephrotoxicity rates between 15% and 58%^23–28 in the pediatric setting are relevant for physicians choosing the most appropriate antifungal therapy (Table 1).

TABLE 1: Data of Pediatric Amphotericin B Associated Nephrotoxicity

In children, infusion-related adverse effects have been reported less frequently with liposomal amphotericin B and more frequently with amphotericin B colloidal dispersion compared with the deoxycholate formulation. However, there is no consensus on whether (1) lipid-associated amphotericin B formulations are more efficacious than conventional amphotericin B, (2) pharmacoeconomic parameters justify the higher costs of the lipid-associated formulations and (3) salt supplementation and close monitoring of renal function can limit the role of these costly alternatives in non-neonatal pediatric patients.^{1,4,6,23–25,28,30,33}

It is with all these facts in mind that physicians must determine the most cost-effective polyene antimycotic treatment for their pediatric patients. This decision is all the more pertinent in developing countries where resources are scarce. Unfortunately, systematic studies on the nephrotoxic effects of amphotericin B in children are limited.²⁴ The purpose of this article is to review and discuss the literature on deoxycholate amphotericin B-associated nephrotoxicity in the pediatric setting.

A literature search was performed in the PubMed MEDLINE database using the search terms “amphotericin B and nephrotoxicity” and “pharmacokinetics and amphotericin B” with the following filters: Clinical Trial, Review, Comparative Study, Systematic Reviews, Randomized Controlled Trial, Controlled Clinical Trial, Practice Guideline, Meta-analysis and Guidelines, Age: birth to 18 years and the use of human studies. Articles published in English, Spanish or French from January 1, 1959, to December 31, 2012, were reviewed. The reference lists in the reviewed articles were manually searched for additional relevant studies on pediatric amphotericin B-associated nephrotoxicity.

MECHANISMS OF ACTION

Amphotericin B is a lipophilic amphoteric polyene macrolide antibiotic synthesized by actinomycete Streptomyces nodosus.^7,34 The main fungicidal activity of this compound is the result of its binding to ergosterol, the major sterol of fungal cytoplasmic membranes.^5,35 This results in the formation of ion channels that increase permeability to small molecules and, consequently, cell death.^9,35 Furthermore, amphotericin B may act as a proinflammatory agent, stimulating innate host immunity, and its oxidative effects may further enhance its antifungal activity.^9,35,36

NEPHROTOXICITY

In addition to binding to ergosterol, amphotericin B also binds to cholesterol, the main sterol of mammalian cell membranes, albeit with less avidity, resulting in nephrotoxic effects.^35,37 Various mechanisms have been proposed to explain renal toxicity. Ischemic injury is believed to occur through a reduction in the renal blood flow and glomerular filtration rate because of increased arteriolar resistance. Different theories to explain the vascular constrictive effects involve the renin-angiotensin system and tubuloglomerular feedback, the renal sympathetic nervous system, calcium mediated mechanisms, arachidonic acid metabolites and the direct interaction of amphotericin B with renal endothelial and smooth muscle vascular cell membranes.^3,38–42 Sawaya et al^40,41 have proposed that amphotericin B binds to membrane-bound cholesterol in smooth muscle vascular cells and forms ion channels that increase the permeability to sodium. The increase in intracellular sodium results in cell depolarization, which may lead to the opening of voltage-dependent calcium channels. Arachidonic acid metabolism may be activated by the elevated intracellular calcium and result in the accumulation of vasoactive substances, which mediate the vasoconstriction of renal vessels. Additionally, the detergent deoxycholate has been associated with nephrotoxicity.⁴³

Amphotericin B affects both glomerular and tubular functions.^3,5 A decreased glomerular filtration rate is evidenced by increases in serum urea and creatinine concentrations.^3,5 Distal tubulopathy is characterized by hypokalemia due to potassium wasting, metabolic acidosis, hypomagnesemia and loss of urine-concentrating ability.⁵ Decreased production of erythropoietin has also been reported.⁴⁴

A decline in the glomerular filtration rate is common to all reports that describe pediatric amphotericin B nephrotoxicity; the magnitude, however, is difficult to establish. This is partly because of the various definitions that have been used. These definitions include: (1) an increase in serum creatinine of at least 0.4 mg/dL, (2) a doubling of serum creatinine from baseline, (3) an increase to ≥1.5 mg/dL from pretreatment levels, (4) a serum creatinine >1 mg/dL and urine output <1 mL/kg/h, (5) an increase of 1.0 mg/dL in the serum creatinine level from baseline or a 50% decrease in the calculated creatinine clearance from baseline, (6) a decrease of >20% of the calculated baseline creatinine clearance and (7) the pRIFLE criteria (Table 1).^23–31 Moreover, serum creatinine levels may not reflect glomerular filtration rates.^45,46 Finally, no long-term, follow-up studies exist on pediatric amphotericin B kidney compromise.

Risk factors for amphotericin B nephrotoxicity in adults include preexisting renal insufficiency, hypokalemia, hydration status, coadministration of nephrotoxic drugs, administration of large individual and cumulative dosages and the nature of the concomitant chronic medical condition of the patient.^3,5,47,48

In a retrospective study, Goldman et al²⁴ assessed 131 different medications in 90 hospitalized pediatric patients concurrently treated with deoxycholate amphotericin B. The patients also received 10 mL/kg of normal saline, as standard care, to prevent nephrotoxicity. The authors reported an increased risk of amphotericin B nephrotoxicity in children concurrently receiving cyclosporine, nystatin or ciprofloxacin.

In a retrospective case control study that involved 138 non-neonatal pediatric patients treated for candidemia with different amphotericin B formulations, Dutta and Palazzi²⁸ observed an increased risk of amphotericin B-associated nephrotoxicity in patients with elevated baseline creatinine clearance and, surprisingly, a lower risk in patients receiving immunosuppressants, including calcineurin inhibitors. The authors stated, however, that patients receiving immunosuppressive therapy may have been more closely monitored, better hydrated and more likely to have been prescribed amphotericin B lipid-associated formulations.²⁸

Dutta and Palazzi²⁸ also reported increased rates of nephrotoxicity in patients who received deoxycholate amphotericin B compared with patients who were prescribed amphotericin B lipid complex (Table 1). Low numbers of patients receiving liposomal amphotericin precluded comparative analyses (Table 1). Interestingly, the reported nephrotoxicity was quickly reversible. A week after therapy, creatinine clearance values in patients who had suffered nephrotoxicity were found to be similar to unaffected patients.²⁸

Renal impairment associated with deoxycholate amphotericin B was described in 9 of 60 children (15%) in a prospective, observational study undertaken to determine its adverse effects.²³ The authors report no requirement of hemodialysis nor any long-term consequences of the patients’ renal compromise.

Four randomized, prospective, controlled clinical trials have compared the nephrotoxicity of conventional amphotericin B and lipid-associated formulations in pediatric patients with neutropenia that received empiric antifungal treatment for fever of unknown origin (Table 1).^25,27,30,31 Prentice et al²⁵ (204 patients) and Walsh et al³¹ (95 patients, data obtained from Blyth et al²) compared the nephrotoxicity rates between liposomal and deoxycholate amphotericin B-treated children, whereas Sandler et al²⁷ (49 patients) and White et al³⁰ (47 patients) compared the rates of nephrotoxicity in children treated with colloidal dispersion and conventional amphotericin B. The results were discordant; although the differences in the first 2 reports were not statistically significant, those in the latter 2 were. However, possible duplicate reporting of data should be considered when drawing conclusions from the latter 2 trials.

A Cochrane systematic review that pooled the 395 children enrolled in these 4 trials^25,27,30,31 revealed a 57% decreased risk of nephrotoxicity when amphotericin B lipid formulations were administered compared with conventional amphotericin B (relative risk: 0.43; 95% confidence interval: 0.21–0.90; P < 0.02).² A sensitivity analysis performed to investigate missing data also showed decreased risk of nephrotoxicity with lipid-associated formulations with respect to deoxycholate amphotericin B, both in best and worst case scenarios. The number of patients who needed to receive amphotericin B lipid-associated formulations instead of conventional amphotericin B to avoid 1 child developing nephrotoxicity was 6.

However, the authors stated that these results must be interpreted with caution because they observed significant heterogeneity among the studies as well as possible duplicate reporting.² Furthermore, they affirmed that the decreased nephrotoxic effect was not maintained if either Sandler et al²⁷ or White et al³⁰ trials were excluded from the analysis.² These observations are of pharmacoeconomic importance because the excluded studies prescribed amphotericin B colloid dispersion, the least costly of the 3 lipid-associated formulations and the 1 with the most serious infusion-associated side effects.^21,27,30

Some authors have reported misgivings about the overall conclusions of Blyth et al⁵⁴ regarding “optimal treatment strategies” due to methodological issues. Moreover, heterogeneity was not considered when Blyth et al² combined Walsh’s and Sandler’s studies with either White’s or Prentice’s. Nevertheless, we believe that most of the methodological issues do not apply to the particular analyses of nephrotoxicity. Therefore, we consider the lack of significant difference in nephrotoxicity rates between liposomal and deoxycholate amphotericin B to be valid.

PHARMACOKINETICS

Because of its insolubility in water, amphotericin B is combined with the detergent deoxycholate in a 3:7 compound for intravenous use.^1,34,55 In the bloodstream, it dissociates from the carrier and is distributed with lipoproteins, being preferentially taken up by organs of the reticuloendothelial system (ie, liver, spleen, lungs, bone marrow), lungs and kidneys.^1,7,35 Amphotericin B then slowly reenters circulation.^1,9

In general, the organs principally responsible for drug pharmacokinetics are the liver and the kidneys.^56,57 However, in the case of amphotericin B, only a small percentage of the parent drug molecule is excreted in urine or bile and no metabolites have been identified. Therefore, tissue accumulation is believed to account for most of its deposition.^37,58 Developmental changes in toddlers’ renal function, including “overshoot” of the glomerular filtration rate above levels encountered in older children and adults,^57,59 would therefore not explain the dissimilar deoxycholate amphotericin B pharmacokinetics and ensuing lower nephrotoxicity in younger children. Blood concentrations do not vary with kidney or liver failure. Different pharmacokinetics of amphotericin B have been reported to account for the milder nephrotoxicity observed in children than adults.^7,24,27,49 However, a closer look at the data^{16,49–53,60} not only calls this observation into question but also demonstrates that pediatric pharmacokinetic parameters vary considerably, particularly among premature infants (Table 2).

TABLE 2: Pharmacokinetic Parameters of Deoxycholate Amphotericin B

The half-life of amphotericin B in the initial phases of treatment has been reported to be shorter than in the later phases,⁵² indicating a possible accumulation of the drug and achievement of a steady state after a few days of therapy.^52,53 Plasma amphotericin B concentrations also have been shown to be 10-fold higher than accepted antimycotic levels after 7–18 days of therapy, further supporting the hypothesis of possible drug accumulation.⁵⁰ However, the reported peak plasma concentrations after a 1 mg/kg dose also vary widely (0.54–4 μg/mL), most likely related to the patients’ age, phase of treatment and the individual pharmacokinetic parameters of the studied patients.^{34,49–51,53,60}

In the pediatric setting, the half-life of amphotericin B is not uniform.^{16,49–51,53} An inverse correlation between patient age and half-life has been reported.⁵⁰ However, this observation is unexpected as it contrasts with the more rapid amphotericin B clearance rates found in infants compared with older children,⁵¹ and other results in which lower amphotericin B exposure was observed in lighter, younger patients with respect to older heavier ones.⁶⁰

Clearance rates of amphotericin B, when corrected for body weight, have been reported to be more rapid in children than in adults (mean 3.36 ± 3.86 vs. 0.43 ± 0.08 mL/min/kg).^49,52 This is consistent with the lower serum levels reported in children when given comparable doses.⁴⁹ These results, however, have been contested; other authors described adult amphotericin B clearance rates similar to those found in children (mean 0.46 ± 0.20 and 0.43 ± 0.08 vs. 0.43 ± 0.08 mL/min/kg; Table 2).^50–53 The results of reported pediatric distribution volumes also show variation; some studies show similar values in adults, whereas others have dissimilar volumes of distribution (Table 2).^16,50–53

When considering these conflicting pharmacokinetic data, methodological differences should be taken into account. These include different pediatric ages, variable analytical methods used to determine amphotericin B concentrations, small sizes of reported series, single or multiple dosing studies and the lack of statistical analysis when comparisons were made.^{16,49–53,60} Nevertheless, 5 conclusions may be drawn: (1) there is marked variability in the pediatric pharmacokinetic parameters of amphotericin B, particularly in very premature neonates; (2) the assertion that the milder nephrotoxicity in children compared with adults is because of different amphotericin B pharmacokinetics needs to be revised; (3) amphotericin B exposure differs in older, heavier children compared with younger, lower weight patients, implying overdosing in the older ones (with the possibility of nephrotoxicity) and underdosing in the younger ones (posing greater risk for treatment failure); (4) pediatric patients tend to achieve a steady state after a few days of starting antimycotic therapy, with initially shorter half-lives progressing toward longer ones with higher concentrations, possibly reflecting an accumulation of amphotericin B^{16,49–51,53} and (5) these diverse pharmacokinetic data of conventional amphotericin B preclude comparison with pharmacokinetic data of lipid-associated formulations in the pediatric setting.

STRATEGIES FOR MITIGATING DEOXYCHOLATE AMPHOTERICIN B-ASSOCIATED NEPHROTOXICITY IN THE PEDIATRIC SETTING

Sodium depletion has been reported to enhance the toxic effects of amphotericin B on glomerular function in adults and, in an experimental setting, in anesthetized dogs. Function improved with sodium correction.^61–63 In the only prospective, double-blind, placebo controlled trial that studied sodium loading effects, Llanos et al⁶⁴ concluded that saline infusion, but not electrolyte-free water infusion, protected glomerular function in adults at low risk for nephrotoxicity who were being treated with amphotericin B for leishmaniasis infections.

A high Na+ intake (>4 mEq/kg/d) has been shown to lessen amphotericin B-induced nephrotoxicity in 2 small retrospective studies of extremely premature infants and was recommended by the authors as an adjunct treatment when this drug is prescribed.^32,65

The authors further stated that ethical equipoise precluded them from a prospective randomized study in which patients would have been assigned patients to a low Na+ intake group.⁶⁵ However, the results must be analyzed with caution as both studies contained duplicate reporting of patients.

Turcu et al⁶⁵ speculate that sodium loading inhibits tubuloglomerular feedback-induced activation by amphotericin B, thereby improving renal blood flow and subsequent ischemia.³ In their study, high fluid intake did not influence amphotericin B nephrotoxic effects in premature infants, which is consistent with Llanos et al’s⁶⁴ report on adult patients.^32,65 In contrast, Sawaya et al⁴¹ have proposed that as a consequence of acute sodium loading, volume expansion leads to increased plasma levels of atrial natriuretic peptide, which would activate intracellular cyclic nucleotides that antagonize the effects of amphotericin B.

Other authors recommend an infusion of 10–15 mL/kg of normal saline (1.5–2.3 mEq/kg of sodium) as standard care to prevent deoxycholate amphotericin B nephrotoxicity in children.^5,24 To date, the exact amount of sodium required to prevent conventional amphotericin B nephrotoxicity has not been established. However, a study reporting that a sodium intake between 3 and 4 mEq/kg/d was somewhat beneficial in preventing renal compromise but that a sodium intake >4 mEq/kg/d prevented nephrotoxicity³² suggests that higher sodium intakes are more effective as a protective strategy.

Amphotericin B lipid-associated formulations were approved by the FDA for antimycotic systemic treatment in the 1990s, providing a higher therapeutic-to-toxic ratio than conventional amphotericin B.²¹ In contrast to conventional amphotericin B, the carriers of lipid-associated formulations are composed of membranes of lipid molecules with hydrophilic heads that face outward to shield the hydrophobic tails. Within these membranes, amphotericin B becomes soluble in plasma and available for distribution.³⁷ Although they are also taken up by organs of the reticuloendothelial system, these formulations functionally spare the kidneys.³⁷ Different theories that may explain the reduced nephrotoxicity include the selective transferring of the drug to the fungal cells and the greater affinity that the synthetic phospholipids have for amphotericin B compared with the cholesterol of the renal epithelial cell membranes.^19,35 Given this greater affinity, higher doses of lipid-associated formulations are required for the delivery of sufficient amounts of amphotericin B to the infection sites.³⁵

Judicious use of nephrotoxic drugs, such as vancomycin, aminoglycosides, calcineurin inhibitors and their combinations, together with adequate hydration have been recommended as important measures for minimizing nephrotoxicity.³⁸ It has been proposed that both clinical and laboratory assessments, which include fractional sodium excretion, urine sodium and osmolality or specific gravity, be performed for the evaluation of nonobvious sodium depletion, unapparent hypovolemia and diminished renal perfusion, before commencing amphotericin B therapy.³⁸ In dehydrated patients, recommendations also include good fluid management and electrolyte replacement therapy to establish adequate renal perfusion.³⁸

Goldman and Koren⁵ have suggested a stepwise approach for the prevention of amphotericin B nephrotoxicity that comprises an assessment of the concomitant use of nephrotoxic drugs and preventive measures that include the infusion of normal saline and consideration of lipid formulations of amphotericin B for (1) children with infections that are unresponsive to deoxycholate amphotericin B or with known side effects to the drug, (2) children concomitantly receiving other nephrotoxic drugs and (3) children with reduced glomerular filtration rates while on conventional amphotericin B. In a later report, Goldman et al further suggested that physicians should consider the use of deoxycholate amphotericin B only for patients with proven fungal infections (clinically or based on cultures), while limiting its application as a prophylactic agent.²⁴ Presently, there is no label for the prophylactic use of amphotericin B.

Alternative antimycotic species-specific drug treatments with a safer kidney profile have been developed. Selected azoles and echinocandins are currently recommended treatment modalities for certain pediatric-invasive mycoses.^35,66–71 Azoles, however, are not free from side effects; important interactions with concomitantly administered agents can either lead to overdosing, causing toxicity, or to underdosing, resulting in a loss of efficacy of coadministered compounds.⁷² Recently, resistance to these agents has also been described.⁷¹

CURRENT RECOMMENDATIONS FOR THE USE OF THE DIFFERENT AMPHOTERICIN B FORMULATIONS IN THE PEDIATRIC SETTING

In neonates, deoxycholate amphotericin B with sodium supplementation and close monitoring of renal function is recommended over lipid-associated formulations, unless significant nephrotoxicity develops.^{1,32,65,68,70} Beyond this age, many recommendations, including alternative antifungal agents, doses and dosing intervals, are extrapolated from adult studies.^{37,50,68,69,71,73}

In a Cochrane systematic review, despite their cautionary note regarding nephrotoxic results, Blyth et al² concluded that, if cost permits, amphotericin B lipid-associated formulations should be used in children instead of conventional amphotericin B. The Infectious Diseases Society of America’s 2009 Clinical Practice Guidelines for the Management of Candidiasis also recommend that deoxycholate amphotericin B should no longer be a first-line therapy option beyond the neonatal period unless other safer drugs are unavailable.⁶⁶ Groll and Walsh³⁷ and Groll and Tragiannidis⁷¹ suggest that, given the available alternatives, only neonates needing antifungal treatment and children with induction therapy for cryptococcal meningitis should be prescribed deoxycholate amphotericin B.

In Europe, the 2012 ESCMID guideline for the diagnosis and management of Candida diseases recommends alternatives to deoxycholate amphotericin B but approves it if the higher toxicity is tolerable from a clinical perspective, with the exception of autologous hematopoietic stem cell transplanted patients.⁶⁸ Nonetheless, the Pediatric Group of the 4th European Conference on Infections in Leukemia does not mention conventional amphotericin B as an option for antifungal treatment in pediatric patients.⁶⁹

This contrasts with Allen et al⁷⁴, who affirm that conventional amphotericin B is an acceptable first-line treatment for persistent febrile neutropenic pediatric patients, leaving lipid-associated amphotericin B for children who are either intolerant or expected to be intolerant to the former pharmaceutical presentation. Horwitz et al²³ also consider that conventional amphotericin B prescription does not significantly compromise patient safety and is appropriate from clinical, economic and ethical perspectives. The authors believe that costly alternative antifungal drug regimens should be reserved for situations where these drugs have been shown to be clinically superior to amphotericin B, such as invasive aspergillosis or in patients with existing renal impairment.²³

In the immediate post-transplant period, antifungal prophylaxis is controversial. Calcineurin inhibitors have been associated with nephrotoxicity in pediatric patients concomitantly treated with deoxycholate amphotericin B.²⁴ Primarily for that reason, lipid-associated amphotericin B, rather than conventional amphotericin B, is recommended if polyene prophylaxis is considered.^67,75 The same holds true for polyene antimycotic treatment; lipid-associated formulations are recommended.^66,75

The FDA has approved the use of lipid formulations of amphotericin B (1) for pediatric patients with systemic mycotic infections, (2) as empiric therapy for neutropenic patients who have persistent fever despite broad spectrum antibiotic therapy and (3) for patients with primary invasive aspergillosis who are intolerant or refractory to conventional amphotericin B, defined as (1) development of renal dysfunction (serum creatinine > 1.5 mg/dL) during antifungal therapy, (2) severe or persistent infusion-related adverse events despite premedication or co-medication regimens and (3) disease progression after >10 mg/kg of accumulated conventional amphotericin B.^33,76 Amphotericin B colloid dispersion has been approved for treatment of proven or probable invasive aspergillosis refractory to or intolerant of deoxycholate amphotericin B at a daily dose of 3–4 mg/kg. The FDA also approved lipid complex amphotericin B at a daily dose of 5 mg/kg for treatment of invasive fungal infections refractory to or intolerant of deoxycholate amphotericin B. Liposomal amphotericin B has FDA approval for empiric therapy of febrile neutropenic patients at 3 mg/kg/d, at a daily dose of 5 mg/kg for patients with invasive fungal infections refractory to or intolerant of deoxycholate amphotericin B and at 6 mg/kg/d for patients with cryptococcal meningitis.³⁷ Many guidelines currently recommend species-specific agents such as voriconazol for aspergillosis and fluconazol for candida, whereas liposomal amphotericin B is suggested as an alternative agent.^35,67–70

DISCUSSION

Although deoxycholate amphotericin B has been the mainstay treatment for most of these life-threatening infections for the past 6o years, its severe toxic effects on kidney function limit its use. Amphotericin B-induced nephrotoxicity, although reversible, carries increased risks of morbidity and mortality, as well as greater economic costs. This situation has led various authors to propose reassessing its place as the standard, both as a therapeutic agent for invasive mycosis and as a point of comparison for antifungal agents. It also brings into question its recommendation in the context of newer agents for the treatment of species-specific invasive fungal infections.^11,77

Deoxycholate amphotericin B has been described as less nephrotoxic in children than adults. Nevertheless, the impact of kidney compromise at this age is cause for concern. Nephrotoxicity early in life may have major adverse effects on children and their long-term well-being,⁵ therefore preventing or attenuating drug-induced nephrotoxicity constitutes a rational health policy, from both medical and financial perspectives.

More modern lipid-associated formulations of amphotericin B, are reported to have similar efficacy to conventional amphotericin B, but less nephrotoxicity. It should be noted, however, that when this issue was addressed in 4 pediatric trials that were later pooled into a systematic comparison between deoxycholate amphotericin B and liposomal or colloid dispersion amphotericin B, low numbers of culture-proven infections limited the power of these studies to detect real differences in efficacy.^{2,25,27,30,31} Very high acquisition costs, ranging from 25 to 230 times the cost of conventional amphotericin B treatment,^1,19,21–23 limit the use of the lipid-associated formulations. Additionally, these formulations are not absolutely free of side effects.⁷⁸

Although lipid-associated amphotericin B acquisition expenditures restrict its use, the financial analysis should go beyond the cost of the drug itself. These analyses should also include the pharmacoeconomic costs of renal compromise, including prolonged hospital stays, alternative monitoring and outpatient follow up of nephrotoxic clinical conditions.^10,13,22,23 Taking the expense of these adverse effects into consideration allows for a more accurate comparison of the treatment costs of the 4 formulations. Unfortunately, potential drug-associated morbidity costs are difficult to consider in financially strained regions where immediate budgeting concerns often override big-picture policies. Nevertheless, compared with deoxycholate, liposomal amphotericin B would not prove cost-effective unless its acquisition costs were substantially decreased.^13,23,77

At present, most data on pediatric-invasive fungal infections, including dosing regimens, have been extrapolated from adult populations.^68,69,72,73 Although the reported prevalence rates for these infections range from 5.3% to 8.2% in febrile neutropenic children,^73,79,80 they are grave medical conditions with major implications in terms of morbidity and mortality. However, for every febrile neutropenic child treated for culture-proven invasive mycotic infection, between 12 and 20 children receive antifungal agents without being infected.

Although there is consensus regarding the use of conventional amphotericin B in neonates, lipid-associated formulations are generally recommended in infants and children when polyene antifungal therapy is considered. With all these contested data in mind, pediatricians must determine the most cost-effective, antifungal treatment for their patients. This is all the more relevant in resource-poor countries.

Identifying patients’ risk factors for invasive fungal infection and nephrotoxicity^24,28,73 are strategies that may modulate deoxycholate amphotericin B prescription and consequently decrease kidney compromise.

Although the “pediatric age” is not a homogenous group, reports that sodium supplementation prevents amphotericin B-associated nephrotoxicity in very premature neonates could support its more widespread use in children when this antifungal is prescribed. The appropriate sodium intake still needs to be established.

Notably, to prevent hyponatremia and its most serious complication, hyponatremic encephalopathy, in hospitalized children, various authors recommend maintenance fluids of half-normal or normal saline solutions (ie, sodium concentration = 77 or 154 mEq/L), as opposed to more hypotonic ones.^81–85 The fluid maintenance requirements for children weighing 10, 20 and 50 kg are 100, 75 and 40 mL/kg/d, respectively.^81,85 As Table 3 shows, in most of the children with these fluid maintenance requirements, the calculated daily Na+ intakes surpass the recommended 4 mEq/kg/d in neonates to protect kidney function when conventional amphotericin B is prescribed. In the future, further studies will be needed to determine whether this higher sodium concentration in maintenance fluids also reduces deoxycholate amphotericin B nephrotoxicity in children.

TABLE 3: Daily Sodium Intake in Children With Different Weights According to Recommended Sodium Chloride Concentration in Maintenance Fluids81–85

Clinical situations encountered in pediatric patients with invasive fungal infections include fever, neutropenia, hypertension, the concomitant use of nephrotoxic drugs such as calcineurin inhibitors and salt retaining medical conditions, including cirrhosis, ascites, hypoalbuminemia, renal hypoperfusion states, heart failure and steroid use.

Pharmacokinetics of deoxycholate amphotericin B could be of use to guide treatment in patients whose medical conditions preclude high sodium intake coadjutant therapy. Unfortunately, to date, the available pharmacokinetic studies of conventional amphotericin B are insufficient to predict a regimen that will prove effective while mitigating nephrotoxicity for any individual pediatric patient. We believe this is because of studies that compare across pediatric age groups and include highly variable neonatal data.

Drugs such as amphotericin B, which have pharmacodynamic antifungal time-course activity characterized by concentration-dependent killing and prolonged post antibiotic effects, are reported to show optimal efficacy when administered in large, infrequent doses.⁸⁶ Combined with the fact that pediatric patients achieve a steady state after a few days of deoxycholate amphotericin B treatment, these findings theoretically would allow for increased initial dosing to achieve rapid fungicidal concentrations, with longer subsequent intervals to minimize dose-related nephrotoxicity.^{16,49–51,53,59} Moreover, studies reporting that amphotericin B exposure differs between older, heavier children and younger, lighter ones⁶⁰ suggest that further research into the pharmacodynamic and pharmacokinetic profiles of conventional amphotericin B may be of use in predicting kidney-sparing, low-cost, effective dosing regimens.

We believe that in light of the current pharmacokinetic data, there are 2 factors that may help explain the reduced conventional amphotericin B-associated nephrotoxicity reported in children compared with adults. The first is the better hydration and higher sodium intake in younger patients. A 70 kg patient administered daily maintenance fluids of 1500 mL/m², even with a hypotonic 40 mEq/L sodium concentration, receives 38 mL/kg/d and 1.5 mEq/kg/d of sodium, whereas a 10 kg child administered maintenance fluids of 100 mL/kg with the same concentration receives a much higher 4 mEq/kg/d sodium intake. The second factor is that a 1 mg/kg prescription of conventional amphotericin B administered to a 50 kg patient represents a 34 mg/m² dose, whereas an 8 kg child receives 20 mg/m².

Upon analyzing the above data, we believe that when polyene antimycotic treatment is necessary, different amphotericin B formulations should be chosen according to each child’s clinical status and risk factors for nephrotoxicity and invasive fungal infection. Low-cost, conventional amphotericin B, with close monitoring of renal function, still has a place in the treatment of empiric invasive mycotic infections for most pediatric patients beyond the neonatal age. Protective measures for kidney function that include the maintenance of effective renal perfusion, sodium supplementation and minimization of concomitant nephrotoxic drugs should be implemented as well. Culture-proven infections should be treated according to antifungal susceptibility, renal function and clinical condition; consideration should be given to the fact that most candida and aspergillosis infections can be effectively treated by fluconazol and voriconazol, respectively.

Although the trials reported by Prentice et al²⁵ and Walsh et al³¹ separately or pooled, failed to show significant difference in nephrotoxicity risks, we concur with Koren et al’s⁸⁷ statement that, “one should err on the side of caution” when polyene antifungal therapy is prescribed. Clinical experience shows that select patients, such as those receiving long-term nephrotoxic agents such as calcineurin inhibitors or those with altered renal function or disease, should begin empiric antifungal treatment with amphotericin B lipid-associated formulations. Although beyond scope of this study, further indications for the lipid-associated formulations, with the exception of amphotericin B colloidal dispersion, should include conventional amphotericin B infusion-related reactions unresponsive to medication.

CONCLUSIONS

We conclude that deoxycholate amphotericin B with close monitoring of renal function can be prescribed to most pediatric patients beyond neonatal age. Select patients, such as those receiving long-term nephrotoxic agents, including calcineurin inhibitors, or those with altered renal function or renal disease, should be treated with liposomal or lipid complex amphotericin B. Regarding efficacy, studies in non-neonatal pediatric patients with culture-proven infections are insufficient to draw conclusions regarding significant differences between the lipid-associated amphotericin B formulations and deoxycholate amphotericin B.

There are few pharmacoeconomic studies that involve liposomal and lipid complex amphotericin B in children. However, unless purchasing prices of the former are substantially reduced, the important resources required for overall substitution of deoxycholate amphotericin B by the lipid-associated formulations may be better allocated to more clinically justified circumstances.

With regard to measures that can be taken to support kidney function, concomitant sodium supplementation is a low-cost option that has been shown to reduce deoxycholate amphotericin B nephrotoxicity in neonates. We speculate that further pharmacokinetic and pharmacodynamic studies in the pediatric setting may also offer physicians appropriate conventional amphotericin B regimens with reduced nephrotoxicity in children. In an era when health cost-effectiveness is an increasingly relevant issue, additional assessment of these renal sparing strategies is imperative to include low-cost deoxycholate amphotericin B in clinical practice guidelines for the management of pediatric fungal infections.

ACKNOWLEDGMENTS

The authors thank Ms. Charlotte Jenkins, Carla Bes and Beatrice Murch for their helpful assistance in editing the manuscript.