FormalPara Key Points

Antifungal pharmacokinetics in children frequently differs from adults, necessitating dosing adjustment to match adult exposure.

Antifungal target plasma concentrations are still largely extrapolated from adult data; therefore, pharmacokinetics/pharmacodynamics indices specific to children must be better defined.

1 Introduction

Invasive fungal diseases (IFD) cause significant mortality and morbidity in children. Candida sp. infect more commonly children than adults and are a major cause of IFD [1] with a high mortality of 34 % in extremely-low-birth-weight infants [2]. Invasive aspergillosis (IA) is another important cause of IFD in immunocompromised children, resulting in unacceptably high mortality despite antifungal therapy (nearly 50 %) [3].

Therapeutic options for IFD are evolving, and several antifungal classes are available to clinicians. Pharmacokinetics (PK), pharmacodynamics (PD), and safety data are predominantly available in adults. However, IFD pathophysiology may differ in children; for example, one of the characteristics of neonatal candidiasis is the high frequency of meningoencephalitis reported in 8–28 % of neonates with invasive candidiasis, likely because of an immature immune system and a more permeable blood-brain barrier [46]. This incidence, however, is under-reported given the difficulties in growing Candida in the microbiology laboratory and lack of available brain tissue samples for culture. In addition to pathophysiology differences, changes in PK in children compared with those in adults might lead to suboptimal drug exposure or an increase in toxicity. In this review, we critically examined the literature on PK and PD of systemic antifungal agents in the pediatric population. A better understanding of these pharmacological concepts will help optimize and personalize antifungal therapy in children and identify areas of future research. The following sources were searched: MEDLINE, clinicaltrials.gov, dailymed.nlm.nih.gov, ema.europa.eu, and international proceedings and abstracts from the earliest record to 15 November 2013. The search strategy included the following key words: ‘pharmacokinetics’, ‘pharmacodynamics’, ‘antifungal’, ‘amphotericin B (AmB)’, ‘liposomal amphotericin B (L-AmB)’, ‘deoxycholate amphotericin B’, ‘amphotericin B lipid complex (ABLC)’, ‘amphotericin B colloidal dispersion (ABCD)’, ‘fluconazole’, ‘itraconaozle’, ‘voriconazole’, ‘posaconazole’, ‘triazoles’, ‘ravuconazole’, ‘isavuconazole’, ‘albaconazole’, ‘echinocandins’, ‘micafungin’, ‘caspofungin’, ‘anidulafungin’, ‘aminocandin’, ‘5-flucytosine’, ‘flucytosine’, ‘children’, ‘infants’, and ‘neonates’. Electronic searches were supplemented by hand searching the reference lists of previous systematic reviews. The search was restricted to trials published in English.

2 Polyenes

The polyene macrolide class includes AmB and newer lipid-based formulations: ABLC, ABCD, and L-AmB. Although these agents are among the oldest class of antifungals and are associated with frequent toxicity, they still play a major role in the treatment of IFD in children [7]. The biggest advantage of polyenes is their wide spectrum of activity.

2.1 Amphotericin B Deoxycholate

By binding ergosterol, a component of the fungus cell wall, AmB increases membrane permeability and induces cell death [8, 9]. In vitro fungicidal activity has been demonstrated against a wide spectrum of fungi, including Candida sp., Aspergillus sp., Zygomycetes, and dimorphic fungi [10]. Although occasional resistance has been reported for multiple fungi species, clinically significant resistance is rare and is mostly encountered in A. terreus, C. lusitaniae, Trichosporon sp., Scedosporium sp., and Fusarium sp. [1114].

AmB activity is concentration dependent with a prolonged post-antifungal effect (PAFE >12 h) [15]. In vitro, its fungicidal effect was observed when the maximum concentration (C max) was four times greater than the minimum inhibitory concentration (MIC) for C. albicans, whereas in candidiasis murine models, maximal activity occurred at C max:MIC of 2.4 [10, 16, 17]. In vitro susceptibility testing has been correlated with clinical outcome in adults with IA, but no correlation was established in adults with invasive candidiasis [18, 19]. No PK/PD indices are validated in children.

Because of poor oral absorption [20], systemic AmB is administered parenterally. AmB circulates in plasma highly bound to lipoproteins and is taken up by reticulo-endothelial organs, especially the liver [21]. In adults, cerebrospinal fluid (CSF) penetration is limited (2–4 % of serum concentrations) [22]. Immaturity in the blood-brain barrier can lead to enhanced central nervous system (CNS) penetration in premature infants where CSF:plasma ratios can reach 40–90 % [23]. AmB is slowly excreted in urine and bile, and metabolism pathways are not well defined [21]. PK in children is characterized by a lower volume of distribution (V) (0.4–3.1 L/kg) and faster clearance (CL) (0.03–0.22 L/kg/h) compared with adults (V 4 L/kg; CL 0.03 L/kg/h) (Table 1) [2327]. Although most PK studies in children support a dosage regimen of 0.5–1 mg/kg/day [2325, 27], a population PK analysis suggested that younger children might be underexposed with 1 mg/kg/day while older children may be overdosed at the same regimen [26].

Table 1 Summary of pharmacokinetic and pharmacodynamic parameters

Similar to adults, AmB use in children is associated with infusion-related toxicities (fever, chills rigors), nephrotoxicity, and electrolyte disturbances [25]. AmB is often believed to be better tolerated in children than in adults because of a decreased number of co-morbidities and co-medication. However, in infants, renal toxicity rates vary across studies from 0 % to 44 % whereas a rate of 53 % was described in adults with IA [2831]. Conversely, up to 50 % of adults experience fever during infusions whereas this toxicity is rarely described in infants [28, 30, 32].

2.2 Lipid-Based Amphotericin B Preparations

Lipid-associated formulations have the same mechanism of action and antifungal spectrum as AmB, but higher dosages are required for equivalent antifungal efficacy in vitro and in animal models [33, 34]. In a murine candidiasis model, AmB was 5- to 8-fold more potent than L-Amb and ABLC in the kidneys [33]. Nonetheless, in adults with hematologic malignancies, no difference in therapeutic efficacy has been demonstrated between the two formulations [35, 36]. In a cohort of children (median age 7 years) with proven IFD caused by C. albicans, Aspergillus sp., or Scedosporium sp., C max:MIC was statistically higher in children with a complete response compared with children with a partial response (median C max:MIC of 68 vs. 40, p = 0.02) [37]. Other than in this small cohort, no further correlation between PK and PD has been established in children.

Compared with AmB, the newer lipid-based formulations have different pharmacological properties. ABLC at a dosage of 2.5 mg/kg/day showed a lower C max (2.1 mg/L) and faster CL (0.218 L/kg/h) than conventional AmB in children with invasive candidiasis (C max of 2.9 mg/L and CL of 0.040 L/kg/h after dosing of 1–1.5 mg/kg/day of AmB) [24, 38]. ABLC was also investigated in neonates with invasive candidiasis [age 8–89 days; gestational age (GA) of 24–41 weeks], showing a clearance rate at the upper range of adults values (0.4 L/kg/h) [39, 40]. This neonatal population analysis led to dosing recommendations of 2.5–5 mg/kg/day (Table 2). In contrast to ABLC, a population PK study of L-AmB in 39 children (aged 0.2–17 years) revealed a higher C max (11.4–44.2 mg/L) compared with similar AmB (Table 1) [37]. In this study, L-AmB also had a lower V and CL, presumably because of a slower tissue distribution [37]. PK estimates for each of these lipid-based formulations were nonetheless comparable to adult values. No published data were found on ABCD PK in children.

Table 2 Dosing recommendations of antifungal agents in children

There have been concerns that amphotericin lipid formulations penetrate kidneys and the CNS to a lesser extent than AmB [41]. For infants, this characteristic is of special interest because Candida commonly disseminates in those two compartments. In animals, Candida kidney infection models showed decreased penetration and potency of lipid formulations compared with AmB [33, 41]. Despite these findings, neonates treated with ABLC demonstrated urine amphotericin concentrations higher than the MIC of many Candida isolates [40]. In the CNS, studies in animal models demonstrated that AmB and L-AmB had higher penetration than ABLC and ABCD, consistent with clinical findings in which neonates treated with ABLC had low or undetectable amphotericin concentrations in CSF [40, 42]. The clinical significance of these findings is not well established. The largest comparative effectiveness study between AmB and lipid-formulation in infants consisted of a cohort study (730 infants, <120 days old) in which there was increased mortality and therapeutic failure with lipid formulations compared with AmB [43]. These findings are consistent with lower lipid products penetration to the kidneys and CNS in animal models, but this observational study failed to adjust for renal function and other clinical data and these results could also be due to confounding by indication.

In terms of safety, the lipid-based formulations offer the advantage of reduced toxicity compared with AmB, with the exception of ABCD for which the frequency of infusion-related symptoms in adults is similar [44]. In infants, uncontrolled studies demonstrated favorable safety for L-Amb with mild increases in liver enzymes (0–37 %) and serum creatinine (0–5 %), and a decrease in potassium (0–5 %) [4548]. No serious AEs were reported in those studies. For ABLC, two large safety studies were conducted in children [49, 50]. The first study involved 111 children (21 days to 16 years) who received ABLC at a dosing of 5 mg/kg/day and demonstrated no drug-related adverse events (AEs) [49]. The second involved 548 children (0–20 years) who received 5 mg/kg/day of ABLC and showed an increase of serum creatinine of 2.5× baseline in 8.8 % of children [49, 50].

Given the extensive clinical experience with AmB and broad-spectrum activity, amphotericin B compounds are widely used for IFD therapy in pediatric populations. Lipid-based formulation PK data are limited in infants, and there are concerns of insufficient penetration in the urinary tract and the CNS in this population [43]. As a result, AmB is generally the first-line polyene agent in young infants unless urine infection and meningoencephalitis are excluded. Further comparative data on the efficacy and safety of AmB vs. lipid-based formulations are needed in infants. Appropriate plasma concentrations have not been defined for safety and efficacy, and therapeutic drug monitoring (TDM) is therefore not recommended.

3 Triazoles

Triazole agents inhibit the cytochrome P450 (CYP) that synthesizes ergosterol, which is a major cell membrane component of most fungi [51]. This mechanism of action inhibits cell growth and replication. In humans, because triazole agents are inhibitors of CYP enzymes (particularly CYP3A4), this class is prone to significant drug-drug interactions. Despite this characteristic, clinical efficacy of triazoles has been demonstrated for the prophylaxis and treatment of many IFDs.

3.1 First-Generation Triazoles

Fluconazole and itraconazole constitute the first generation in this class. They are active against Candida sp., Cryptococcus neoformans, and dimorphic fungi. Exceptions include C. glabrata and krusei, which are often resistant to fluconazole (MIC90 of 32 mg/L and 64 mg/L, respectively) [52]. Only itraconazole provides coverage of Aspergillus sp. Both agents are available as oral and parenteral formulations.

3.1.1 Fluconazole

Fluconazole exhibits in vitro fungistatic activity [53], which does not correlate well with concentration and is more dependent on duration of exposure (time dependent). In vivo infection models demonstrated that a ratio of area under the concentration-time curve (AUC) over MIC (AUC:MIC) of 18 was associated with 80 % survival in the murine candidiasis models [54, 55]. In adults with Candida infection, a ratio (AUC:MIC) <11.5, and MIC ≤64 mg/L were associated with increased mortality or therapeutic failure [56, 57]. For a Candida sp. with an MIC of 32 mg/L, AUC:MIC >12 corresponds to an AUC0–24 of 400 mg*h/L, which is often cited as the exposure target for adults. In immunocompromised patients and premature infants, an AUC0–24 of 800 mg*h/L is typically targeted, probably to cover a broader range of Candida sp. MICs. In children, there is no established relationship between PK and PD, and exposure target is usually extrapolated from adults (AUC0–24 from 400 to 800 mg*h/L).

Fluconazole has high oral bioavailability (92 %) [58]. Consistent with its low protein binding, fluconazole shows good penetration into tissues and body fluids, especially in the urine, and concentrations in CSF achieve 80 % of blood levels [59, 60]. Fluconazole does not undergo significant metabolism and is predominantly excreted unchanged in urine. After filtration, fluconazole is highly reabsorbed by the kidney [61]. Dosing should be reduced in subjects with renal impairment; however, continuous renal replacement therapy might result in lower concentrations than expected because of the lack of renal reabsorption in this setting.

In neonates, a population PK model developed with a cohort of 55 infants aged <120 days (23- to 40-week gestation) revealed the importance of gestational age and postnatal age in fluconazole disposition [63]. Based on this PK model, simulations predicted that a daily dose of 12 mg/kg in infants <30 weeks of gestation achieved a median 24-h AUC close to the target of 800 mg*h/L, whereas infants >30 weeks of gestation achieved a lower median 24-h AUC of 400 mg*h/L [63]. In contrast, dosing of 6 mg/kg daily in adults is sufficient to achieve a 24-h AUC of 400 mg*h/L in more than 85 % of the subjects [56]. In the population PK model involving infants described above, fluconazole achieved steady state therapeutic AUC in 5–7 days [63]. As a result, administration of a loading dose has been suggested in vulnerable populations such as infants [7]. In a small cohort of 10 infants aged <60 days, a fluconazole loading dose of 25 mg/kg was safe and achieved the therapeutic target more rapidly than traditional dosing, which may lead to decreased morbidity [64].

Beyond the neonatal period, the PK in children is characterized by a higher CL (0.030 L/kg/h) and a larger V (0.95 L/kg in children aged 2–12 years) compared with adults (CL of 0.016 L/kg/h and V of 0.7 L/kg) as demonstrated in a study involving 100 children (aged 0–18 years) (Table 1) [59, 65]. Consequently, children require a proportionately higher dose to match adult exposures. Finally, although limited, assessment of oral absorption after 2 years of age suggested similar bioavailability as adults (>80 %) [65].

Fluconazole is a potent inhibitor of CYP2C9 and CYP3A4 leading to increased plasma concentration of other drugs metabolized by those metabolic enzymes when co-administered with fluconazole. Such compounds include tacrolimus or cyclosporine, which are commonly used in the pediatric population at high risk for invasive fungal infection [66]. Beside safety concerns related to drug-drug interactions, fluconazole is well tolerated in children as evidenced by a safety analysis in 562 children (aged 0–17 years) following oral and intravenous administration [67]. The most common treatment-related AEs were gastrointestinal symptoms (7.7 %) and skin rash (1.2 %). In this study, overall, 18 of 562 children (3.2 %) discontinued fluconazole because of AEs [67].

Fluconazole is widely used for the prevention and treatment of invasive candidiasis in children and infants. It is also used for the treatment and secondary prophylaxis of cryptococcal meningitis. PK and safety are well defined across a wide age range. Moreover, its reliable penetration into CSF and the urinary tract makes fluconazole a useful agent against neonatal invasive candidiasis [68, 69]. Children generally require higher dosing per unit of weight to match adult exposure. However, target exposure is extrapolated from adults, and age-specific PD indices have not been established.

3.1.2 Itraconazole

Similar to fluconazole, itraconazole is fungistatic against yeast-like fungi [70], while only itraconazole has fungicidal activity against molds [71]. Given the time-dependent PD demonstrated for both yeasts and Aspergillus sp. [70, 72], separating the daily dose in two appears more appropriate (Table 2) [73]. PD modeling revealed that AUC:MIC, C max:MIC, and minimum plasma concentration (Cmin):MIC equally correlated with antifungal efficacy for the treatment of oropharyngeal candidiasis in human immunodeficiency virus (HIV)-infected children [73].

Itraconazole circulates in blood highly protein bound and undergoes extensive hepatic metabolism into several metabolites. One of those metabolites, hydroxyl-itraconazole, has similar activity to the parent drug. Itraconazole is excreted through the liver and kidneys. Oral absorption is variable, but acidic gastric environment, food, and administration in oral solutions enhance its absorption [81]. Dosing of 5 mg/kg/day with an oral solution results in lower exposure in children compared with adults (mean AUC0–24 of 8.7 mg*h/L vs. 22.7 mg*h/L, and mean C max of 0.6 mg/L vs. 1.5 mg/L) [74, 75]. These differences were not found with lower dosing (2.5 mg/kg/dose twice a day), which produced similar exposure in adults and children [76]. These findings highlight the non-linearity of the PK profile of itraconazole and the possible dose-dependent bioavailability. PK parameters in children aged >5 years are otherwise relatively similar to adults [73, 74, 77].

Itraconazole is well tolerated in children, and the most commonly reported AEs are gastrointestinal symptoms (8–12 %) [73, 78]. In adults, trough concentrations above 17 mg/L (measured by bioassay) are significantly associated with toxicity (mainly fluid retention and gastrointestinal symptoms) [79]. The main safety concern for itraconazole is the potential for drug-drug interactions due to inhibition of CYP3A enzymes [80]. Co-medication with itraconazole may result in increased plasma concentration of drugs metabolized by CYP3A4 enzymes, such as cisapride or oral midazolam [81]. Enhanced vincristine neurotoxicity is a well-documented drug interaction with itraconazole in both adults and children [80, 82]. In a retrospective study of 20 children with acute lymphoblastic leukemia, those receiving vincristine in combination with azole treatment (predominantly itraconazole) experienced significantly more peripheral neurotoxicity (p < 0.05) [82].

Itraconazole use is limited in children because of erratic, dose-dependent, oral bioavailability, high PK variability, and the availability of more reliable alternatives [83]. Given established target trough concentrations in adults, TDM is standard practice (target trough concentrations >0.5 mg/L when measured by high-performance liquid chromatography [56] and <17 mg/L when measured by bioassay [79]). No age-specific target concentrations have been identified in children.

3.2 Second-Generation Triazoles

Second-generation triazoles are active against a wide spectrum of clinically important fungi including yeast, molds, and dimorphic fungi. Members of this class of triazoles include voriconazole, posaconazole, and newer compounds such as ravuconazole.

3.2.1 Voriconazole

Voriconazole is available in oral and intravenous formulations and is the primary therapy for IA. It is structurally similar to fluconazole but has extended antifungal spectrum activity against Aspergillus sp. Despite its broad-spectrum activity against yeast and molds, voriconazole is not active against zygomycetes [12, 84]. In vitro, fungicidal, and fungistatic activity against Aspergillus sp. and Candida sp., respectively, are time dependent [85, 86]. Near maximal effect against Candida sp. was observed at concentrations three times the MIC at different time points [86]. In murine candidiasis models, AUC:MIC was strongly predictive of treatment success with a suggested target of free concentration AUC0–24:MIC of 20 [87]. Consistent with the voriconazole time-dependent effect, C min >1–2 mg/L was a good predictor of successful clinical outcome in both adults and children [88, 89]. In children, C min <1 mg/L was associated with increased odds of death (odds ratio [OR] 2.6; 95 % confidence interval [CI] 1.6, 4.8) [88].

Voriconazole is 58 % protein bound, distributes well into tissues and CSF [90, 91], and is extensively metabolized by hepatic CYP2C19. Allelic variations contribute to high inter-subject PK variability in adults and children [92, 93]. In contrast to adults, children have a linear PK at dosing of 3–4 mg/kg intravenously (IV) [94]. Elimination becomes nonlinear over the range of 4–8 mg/kg every 12 h [95]. Children aged <12 years have greater clearance of voriconazole and require almost twice the dose to match adult systemic exposure (7–9 vs. 4 mg/kg IV twice a day) [92, 95, 96]. Children also have a lower oral bioavailability than adults (45–65 % vs. 96 %) [92, 95]. This difference is not completely understood but may relate to greater first-pass metabolism in children.

Voriconazole AEs include visual disturbances, elevated hepatic transaminases, and skin photosensitization (13–30 %) [94, 95, 97, 98]. In adults, trough concentrations above 4–5.5 mg/L correlate with toxicity [99, 100]. In a small pediatric cohort, oral administration of more than 6 mg/kg/dose twice a day was associated with an increased risk of phototoxic skin reactions, but no correlation was described with trough concentrations [98]. For other forms of toxicity, no relationship to exposure has been determined in children [101].

Multiple pediatric PK studies recently helped determine optimal dosing in children down to 1 year of age, but voriconazole currently has European Medicines Agency (EMA) and US Food and Drug Administration (FDA) labeling for children over 2 and 12 years of age, respectively [102, 103]. In infants, PK and safety still need to be characterized, and voriconazole use is discouraged in this population. Efficacy has been well established in adults with aspergillosis, and it is the primary recommended therapy for IA [104]. TDM is recommended, given a high inter-subject variability, and has proven useful in adults for whom a randomized controlled trial showed better outcome and reduced toxicity with target C min between 1 and 5.5 mg/L [105]. In children, C min ≥ 1 mg/L was also retrospectively associated with decreased mortality from IFD, but as opposed to adults, no upper bound of target trough concentration has been identified, and this question warrants further study [88].

3.2.2 Posaconazole

Posaconazole has broad antifungal activity against the majority of yeasts and azole-resistant Candida sp. In addition to Aspergillus sp., posaconazole is also active against other molds, including zygomycetes, as opposed to voriconazole. Similar to other triazoles, posaconazole is fungicidal in vitro with time-dependent killing against most Candida species and molds [62, 106]. In animal models, the duration of exposure to plasma concentrations above the MIC is the most important parameter for optimum efficacy [107].

Posaconazole is only available as an oral formulation and should be taken with high-fat meals to enhance absorption [108]. Protein binding is high, and the primary route of elimination is through feces with renal clearance playing a minor role [109]. Posaconazole prophylaxis and treatment have been described in children, but the PK is not well characterized. No specific dose recommendations exist, and its use in children aged <12 years is off-label [110113]. A twice-daily dosing algorithm based on allometric scaling (body weight based) delivered adequate exposure in 12 children with chronic granulomatous disease (mean trough concentration of 1.54 mg/L) [113]. In adults, the PK is linear, and the long half-life (25 h) produces stable plasma concentrations over time [108]. Steady-state concentrations are achieved 7–10 days after initiating therapy, suggesting that this agent might be suboptimal for induction therapy or a loading dose may be required. Posaconazole is well tolerated, and the most common AE is gastrointestinal symptoms in 25 % of patients [111, 112].

Exposure-response relationship analyses in adults suggest the need for TDM with a target trough concentration of 0.5–1 mg/L for prophylaxis and therapy, respectively [114, 115]. There is no correlation between exposure and toxicity [114].

Posaconazole PK/PD in children is scarce and needs better characterization. It is not licensed for use in children aged <12 years by regulatory agencies [116, 117]. An ongoing trial will address posaconazole dosing in children (NCT01716234). Availability of an intravenous formulation will attract interest from the pediatric community; however, PK studies will need to be conducted to evaluate optimal dosing in children.

3.2.3 Other Triazoles

Ravuconazole, isavuconazole, and albaconazole are the newest antifungal triazoles with in vitro and in vivo fungicidal activity against a wide spectrum of clinically important fungi. They are active against Candida sp. including fluconazole-resistant strains, Cryptococcus sp., and Aspergillus sp. [118122]. Activity against Fusarium sp., Scedosporium sp., and Mucor sp. remains limited [52, 122, 123]. Similar to other triazoles, animal models show that antifungal efficacy is time dependent [118, 124]. All three agents are available orally and have high oral bioavailability, but only ravuconazole and isavuconazole are available intravenously [125, 126]. Clinical experience is still limited in adults, but PK studies reveal a prolonged half-life from 75 to 117 h [125, 128]. These new triazoles are well tolerated, and the most frequently reported adverse effects are headaches, rhinitis, and gastrointestinal symptoms [125, 128]. No clinical trials have been completed in children, and so these agents are not currently recommended.

4 Echinocandins

Agents from the echinocandins class are only available in a parenteral form and include micafungin, caspofungin, anidulafungin, and the more recent aminocandin. They act by inhibiting (1,3)-β-d-glucan synthase, a fungus-specific enzyme crucial to the biosynthesis of glucan in the fungal cell wall [129]. They exhibit fungicidal activity against most Candida sp., including fluconazole-resistant species (C. albicans, C. parapsilosis, C. glabrata, C. tropicalis, and C. krusei) and fungistatic activity against Aspergillus sp. [130, 131]. Echinocandins are not active against C. neoformans, zygomycetes, and dimorphic fungi. Advantages of echinocandins include low toxicity and minimal drug interactions owing to the fungus-specific mechanism of action. Disadvantages include the lack of oral formulation and the reduced activity against non-Candida sp.

4.1 Micafungin

Micafungin displays concentration-dependent fungicidal killing, and in animal models, efficacy correlates best with AUC:MIC ratios [132]. Analysis of adult clinical data for the treatment of invasive candidiasis found that an AUC:MIC ratio over 3,000 predicts better mycological response [133]. In neonates who are at greater risk of meningoencephalitis caused by Candida sp., a different exposure target (AUC0–24 = 166.5 mg*h/L) has been proposed based on a rabbit model of neonatal hematogenous Candida meningoencephalitis [134].

Micafungin is highly bound to protein and distributes well into tissues including the lung, liver, and spleen. Although penetration into the CNS and eye are generally limited in adults, high-dose micafungin successfully treated neonatal animal models of hematogenous Candida meningoencephalitis [134, 135]. Micafungin undergoes only limited phase I metabolism, and the parent drug is mostly excreted via the biliary system [136].

Micafungin PK is well described in pediatric patients aged 4 months to 16 years, and 229 patients were included in population PK analyses in support of recent US pediatric labeling [137]. The half-life is approximately 12 h, and the PK is linear at doses of 1–3 mg/kg [138]. Several PK studies show an age effect demonstrated by an inverse relationship between weight and CL, such that as body weight decreases, higher dosages of micafungin (on a milligram/kilogram basis) are required to achieve equivalent drug exposure [139]. In a study of 77 neutropenic children, doses of 3–4.5 mg/kg/day for 2- to 8-year-olds and 2–3 mg/kg/day for 9- to 17-year-olds achieved adult exposure [140].

Four PK studies demonstrated that younger infants (GA of 24–40 weeks and postnatal age of 2–119 days) have higher clearances and therefore need higher dosing than children and adults [127, 141143]. Moreover, this population requires reliable micafungin CNS penetration because of a higher risk of Candida meningoencephalitis. Based on a population PK model, PK simulation analyses showed that 83 % of neonates receiving 10 mg/kg/day would achieve an AUC0–24 h of 166.5 mg*h/L, a specific target for neonatal Candida meningoencephalitis [134, 141]. These preliminary PK/PD analyses led to the dose selection for an ongoing phase III, randomized controlled trial comparing micafungin with amphotericin B deoxycholate (NCT00815516).

Pooled data from six pediatric studies show that micafungin has a favorable safety profile [144]. The most common treatment-related AEs are transient and include liver enzyme elevations and decreased potassium (3 %) [144]. In a trial comparing micafungin and L-Amb in 106 children with invasive candidiasis, both drugs had equivalent efficacy, but micafungin was better tolerated [145]. No maximum tolerated dose has been identified, and high doses of 15 mg/kg have been safely used in premature neonates [142].

Overall, micafungin dosing is well defined in children, and regulatory agencies have labeled its use in the pediatric population [137, 146]. The EMA product license includes dosing of 2 mg/kg/day for all children aged <16 years, including neonates. However, the FDA product label includes dosing of 2 and 3 mg/kg/day in children weighing <30 kg for the treatment of invasive candidiasis and esophageal candidiasis, respectively. Unlike the EMA, the FDA label restricts its use to children aged ≥4 months. For infants aged ≤3 months, PK studies and age-specific exposure targets suggest that dosing up to 10 mg/kg/day is required. Clinical efficacy and safety data with newly proposed dosing in this population needs further evaluation.

4.2 Caspofungin

Caspofungin is recommended for therapy of invasive candidiasis, salvage therapy of IA, and as empirical therapy in febrile neutropenia in children aged >3 months. It demonstrates in vitro fungicidal activity against Candida sp. with a concentration-dependent effect and prolonged PAFE (12 h) [15, 106, 147]. In vitro PD properties suggest optimal killing when concentration is four times the MIC [147]. Murine candidiasis models demonstrate reduced activity against C. parapsilosis and C. guillermondii. The clinical significance of this reduced activity is unclear, but clinicians should be aware of the possibility of reduced efficacy. Caspofungin is generally classified as fungistatic against Aspergillus sp.; however, evidence of killing in vitro was shown by a change in the cells staining pattern but this phenomenon did not translate into reduction in the number of colony forming units from standard killing curve measurements [148]. In a PK trial of 32 children aged 3 months to 17 years with confirmed or suspected IFD (Candida sp. and Aspergillus sp.), the exposure-response analyses revealed no association with clinical outcome and PK/PD parameters [149].

Caspofungin is metabolized by the liver with a half-life of 7–10 h and demonstrates linear PK in children [150, 151]. Similar to other agents in this class, caspofungin urine concentrations are low [152]. However, this poor renal penetration has not translated into treatment failure of candiduria in adults [153]. Specific to caspofungin, weight-based dosing does not provide consistent steady-state concentrations across ages, whereas dosing based on body surface area does [151]. In infants and children with neutropenia, 50 mg/m2/day dosing produces systemic exposures similar to adults receiving 50 mg/day [150, 151]. In young infants, although the PK has not been fully characterized, a study involving 18 infants aged <3 months with suspected invasive candidiasis showed that 25 mg/m2/day was required to achieve adult plasma concentrations [154].

Dose-limiting toxicities have not been seen in adults [155]. Multiple clinical trials in children found that caspofungin is well tolerated; fever, rash, hypokalemia, and elevated liver enzymes are the most frequently reported AEs but do not usually require therapy discontinuation [149, 151, 154, 156, 157]. The incidence of AEs does not correlate with concentration [149]. In young infants, safety data are limited to two small cohorts (total N = 22) in which AEs were described without a comparator group [158, 159].

Caspofungin is recommended for pediatric IFD and is particularly useful against invasive candidiasis caused by azole-resistant isolates. Its PK has been adequately characterized in children down to 3 months of age. However, PK, safety, and efficacy are inadequately defined in young infants [158, 159].

4.3 Anidulafungin

The spectrum of anidulafungin activity is wide with fungicidal effect against Candida sp. including fluconazole-resistant strains, and fungistatic effect against Aspergillus sp. [119, 160162]. Fungicidal activity was demonstrated in neutropenic murine models for which both C max:MIC and AUC:MIC ratios strongly predicted successful treatment of systemic candidiasis [163]. In Aspergillus animal models, anidulafungin therapy reduces Aspergillus antigenemia but is unable to clear the infection, consistent with its fungistatic activity [119].

Anidulafungin is highly protein bound (>99 %) but achieves tissue concentrations above MIC in animal models [164]. Anidulafungin demonstrates linear PK and a longer half-life than other echinocandins (20 h) [165]. It has a unique elimination pathway among the echinocandins, consisting of slow non-enzymatic degradation to inactive metabolites. When anidulafungin is given at the same weight-adjusted dose (1.5–3 mg/kg loading dose, 0.75–1.5 mg/kg/day maintenance dose), children aged 2–17 years with neutropenia achieve exposures similar to adults [165]. In infants and neonates, 1.5 mg/kg/day produces plasma concentrations comparable to adults [166]. Both prior PK trials demonstrate that anidulafungin is well tolerated with no observed, drug-related serious AEs.

Despite available PK data and a favorable safety profile demonstrated in two pediatric trials, clinical experience in children remains insufficient to recommend it for use in this population. Anidulafungin is not labeled for children aged <16 years in the USA and aged <18 years in Europe [167, 168].

4.4 Aminocandins

Aminocandin is the newest echinocandin and is available as an intravenous formulation. Similarly to other agents of this class, it is active against Candida sp. and Aspergillus sp. in vivo [169, 170]. Fungicidal effect against Candida sp. best correlates with peak:MIC demonstrating a concentration-dependent activity [170]. Animal models suggest that extended dosing interval (7–10 days) is effective as treatment and prophylaxis of invasive candidiasis [171]. In healthy volunteers, a phase I study also demonstrated a long half-life (48–58 h) indicating that dosing could be less frequent than once a day [172]. In this small cohort, aminocandin was well tolerated, but no other clinical trials have been published in adults or children.

4.5 Second-Generation Echinochandins

ASP9726, a novel second-generation echinocandin is under development. As opposed to other echinocandins, ASP9726 showed Aspergillus hyphal growth inhibition and improved MIC against C. parapsilosis and echinocandin-resistant Candida in vitro [173]. To our knowledge, no ASP9726 clinical trials have been registered in clinicaltrials.gov at the time of submission.

5 Nucleoside Analogs

5-Flucytosine (5-FC) is an antimetabolite drug that causes RNA miscoding and inhibits DNA synthesis [174]. 5-FC has activity against Candida sp. and C. neoformans [175, 176]. In vitro and in vivo testing demonstrates that flucytosine is fungistatic against yeasts with concentration-independent pharmacodynamics [177, 178]. Moreover, 5-FC exhibits PAFE up to 10 h [179]. It is never used as monotherapy given the rapid emergence of resistance. Its use is limited for the induction therapy of cryptococcal meningitis for which 5-FC in combination with AmB was shown to be efficacious [180].

Only the oral formulation is available in the USA. Oral bioavailability is high in adults (75–95 %), and a trial comparing IV and oral 5-FC in adults with HIV-associated cryptococcal meningitis did not detect a difference in fungicidal activity [181]. Protein binding is negligible, and distribution into tissues and body fluids is reliable. 5-FC is mainly eliminated in an active form in urine. The elimination half-life is 3–5 h, and administration in three to four daily doses is required [182]. Data on 5-FC PK in children are limited, but drug CL appears slower in children compared with adults [23, 183]; as a result, the adult dose of 100 mg/kg/day might lead to overexposure. In addition, the elimination half-life in neonates is nearly twice as long as in adults (4 vs. 7 h) [23, 183], suggesting that the dosing interval should be longer (8–24 h).

Dose-limiting toxicities described in up to 44 % of adults (hematologic, gastrointestinal, and hepatic) limit the use of 5-FC [184, 185]. Safe target plasma concentrations in adults are established at 40–60 mg/L, but no such data are available for children [184].

Given the narrow therapeutic window, TDM is the standard of care for 5-FC with target peak concentrations of 20–50 mg/L [186]. However, target concentrations have not been established in children. Concentration-independent PD suggests that lower dosages could be considered in future research. Clinical use of 5-FC is limited mainly to combination therapy with AmB for the treatment of cryptococcal meningitis. Because of limited PK, safety, and efficacy data, the use of 5-FC in infants is discouraged.

6 Novel Antifungals Agents under Development

Two antifungals agents are under development and would each represent a new class of antifungals. T2307, a novel arylamidine showed in vitro and in vivo activities against Candida sp., C. neoformans, and Aspergillus species [187].The second agent is kakeromycin, which demonstrated good in vitro activity against the same fungal pathogens than T2307 [188]. For those two potential antifungal drugs, no clinical data are available thus far.

7 Conclusion

Effective and safe antifungal therapy depends on optimal drug dosing. With the recent development of new antifungal therapies, specific data in children are needed because prior extrapolations of adult dosing to children have often proved wrong. Thanks to regulatory initiatives and suitable trial designs, the number of PK studies in children has increased and clarified some dosing issues in this vulnerable population. Based on our review of the literature, we present a summary of antifungal dosing recommendations in Table 2. Whenever dose ranges are presented, clinicians should consider the benefits of early initiation of high-dose therapy in improving outcomes in the setting of severe IFD.

Despite better evidence on antifungal therapy, morbidity and mortality related to IFD remain unacceptably high and dosing should be optimized, especially in infants. Future research should focus on determining drug disposition of newer antifungals such as newer triazoles and echinocandins in children as they become available for adults. Future research questions should also relate to optimizing dosing of older agents such as fluconazole in neonates for which there is a need to better support the use of a loading dose. We also need better comparisons of safety and effectiveness of AmB and lipid formulations of amphotericin B in infants. Finally, clinical validation of PK/PD indices determined in animal models are sparse in humans, even more so in children. Ongoing efforts to characterize the PK and PD of antifungal agents, especially in younger populations, will help inform dosing and improve clinical outcomes of IFD.