FormalPara Key Summary Points

Invasive mold infections, including Aspergillus, Mucorales, and other agents, are becoming increasingly burdensome as survival increases for at-risk patients undergoing treatment for high-risk hematologic malignancies.

Clinicians must maintain heightened awareness for invasive mold infections, particularly as the adoption of antifungal regimens has reduced the incidence of candidiasis.

Much of the literature on treatment of invasive mold infections relies on extrapolation from adult data.

New antifungal agents have been developed which have shown activity against Aspergillus, Mucorales, and other agents, but further research is needed to determine efficacy, safety, pharmacokinetics, and pharmacodynamic targets associated with therapeutic success.

Introduction

The frequency of invasive fungal infections (IFI) has been increasing in recent years, as a reflection of improved survival of patients at risk. New chemotherapy regimens for cancer patients, particularly those with acute myelocytic leukemia (AML), has resulted in increased survival, but these patients experienced prolonged periods of neutropenia. Similarly, new allogeneic human stem cell transplant (A-HSCT) strategies have increased the number of suitable donors through T-cell depletion, use of haplo-identical donors, or use of umbilical cord stem cells, all of which have increased risk of graft versus host disease (GVHD). Increased survival of patients with primary immune deficiencies (PID) contributes to the cohort of patients at risk. A small but important group of pediatric patients at risk for IFI include extreme low birth weight premature newborns < 28 weeks of gestation (ELBW), as well as those who use biological agents for treatment of auto-immune or auto-inflammatory diseases.

Typically, Candida spp. infections have been the most common pathogens responsible for IFI, but recent adoption of infection prevention measures and prophylactic antifungal regimens for at-risk patients have significantly decreased the incidence of candidiasis. Subsequently, this has heightened the awareness of invasive mold infections (IMI), of which Aspergillus spp. infections are the most common, followed by Mucorales (currently accepted term for Zygomycetes) and Fusarium spp. Less common but still important molds are Scedosporium spp. and Lomentospora prolificans. (previously Scedosporium prolificans). This article is based on previously conducted studies and does not contain any new studies with human participants or animals performed by any of the authors.

Epidemiology

Describing the epidemiology of IMIs in children is difficult as the data available have significant limitations. They are frequently retrospective, often single center from different parts of the world, and may span a long period, during which underlying risk factors, diagnostic techniques, and treatment modalities changed overtime. Most include only a defined group, for example children with cancer, or premature infants. A recent international multicenter phase III clinical trial on management of acute lymphocytic leukemia (ALL) identified 233 patients with proven/probable IFI of which 159 were due to molds. These infections occurred significantly more often in older (12–18 years old) patients and those receiving more intense chemotherapy. Mortality 1 year after diagnosis was higher in patients with IFI, and this was more noticeable in patients with IMI (14.5%) compared to Candida (6.1%) [1].

Aspergillus. In a large multicenter (6 large pediatric centers), retrospective (over 5 years) study of proven/probable invasive aspergillosis (IA) cases, Burgos et al. identified 139 cases of IA, with underlying risk factors including malignancy (62.6%), A-HSCT recipients (38.1%), PID (11.4%), solid organ transplant (SOT) recipients (11.4%), or other immune deficiency states (7.1%). Neutropenia (< 500/µL) for ≥ 3 days was present in 59% of patients (30% > 30 days), while 69.1% had received steroids, 90% for ≥ 3 days. GVHD was documented in 30.2% of A-HSCT; overall, 95.7% had ≥ 1 risk factor identified. A. fumigatus (52.8%) was the most frequently identified species followed by A. flavus (15.7%) and A. terreus (4.7%). Most common sites of infection were the lungs (59%), skin (10%), sinuses (9.6%), and central nervous system (CNS) (5.9%). In total, 52.5% of patients died during treatment, of whom A-HSCT recipients carried the highest mortality (OR 6.68 95% CI 2.67–16.21) [2]. Surgical intervention was an independent predictor of survival [2]. In a separate study, Zaoutis conducted a national retrospective cohort study of the epidemiology and outcomes of IA in immunocompromised children. He identified 666 IA patients, for an incidence among immune compromised children (152,231) of 437/100,000 admissions (0.4%). Children with malignancy accounted for 74% of cases, followed by hematological or immunological disorder (28%), A-HSCT (15%), and SOT (1%). Overall, 18% of immune compromised children with IA died compared to 1% without IA (P < 0.001) [3]. Not all cancers have the same risk for IA, as seen in a 5-year single-center observational study of 346 patients receiving treatment for newly diagnosed or recurrent cancer, while not receiving mold-active prophylaxis. All cases (incidence 6.8%) occurred in children with hematological malignancies [4]. Incidence was highest in new or recurrent AML (28%), followed by recurrent ALL (9%) and de novo ALL (2%), emphasizing the role of prolonged neutropenia and treatment with steroids. Data from multiple international studies describe incidence rates of 2.7–8% with a mortality that may reach 80% in some cases [5,6,7,8,9]. ELBW infants are increasingly surviving beyond the immediate neonatal period. In these patients, skin is often the port of entry for localized or invasive infections, which can be associated with significant mortality [10, 11].

Mucorales. Epidemiological data is limited. In this manuscript, we will use mucormycosis to refer to infections caused by members of the order of Mucorales. The genera more often associated with infection (75% of cases of mucormycosis) are Mucor, Rhizomucor, and Rhizopus. Other genera frequently encountered include Lichtheimia (formerly Absidia), Cunninghamella, and Saksenaea. Zaoutis et al. identified 157 pediatric cases of mucormycosis (1939–2004) [12]. Median age was 5 years (IQR 0.16–13) and neonates accounted for 28 (18%) cases, of whom 27 were premature. Neutropenia was as common a risk factor as prematurity. Children with type I diabetes represented 13% (10% had presented with ketoacidosis), while 20 (18%) cases had no identifiable underlying condition. Rhizopus and Mucor were the most common genera identified. The most common sites of infection were cutaneous (27%), gastrointestinal (21%), rhino-cerebral, (18%), and pulmonary (16%). Gastrointestinal cases were significantly associated with prematurity. Cerebral, gastrointestinal, and disseminated mucormycosis were associated with the highest mortality rates (100%, 100%, and 88%, respectively). Surgery was independently protective from mortality [12]. In a subsequent report, Roilides et al. identified 30 new cases between 2004 and 2008. This report included 59 newborns, with a median age of 12 days, of whom 73.8% were ELBW, and the most common patterns of disease were gastrointestinal and cutaneous. The former had a mortality of > 75% [13]. A more recent report identified cases submitted to one or both of two international registries (Zygomyco.net and FungiScope) between 2005 and 2014. Similar to the previous reports, hematological malignancies were the most common underlying conditions and lungs, sinuses/brain, and gastrointestinal tract were the most common sites of infection. The most significant difference between this and the previous reports discussed above was the absence in newborns. It is possible that the way we care for newborns today (central lines instead of arm boards, improved feeding practices) may have led to this difference, considering the previous reports spanned 30 years [14].

Other molds. A recent prospective study conducted by the International Pediatric Fungal Network (12 US and 10 international sites, mainly Western Europe) in children 3 months to 18 years old identified 131 invasive mold infections, of which 75% were Aspergillus spp, 13% were mucormycosis, and 17% were other molds. Among these, Fusarium spp., Curvularia spp. and Exserohilum spp were the most common [5]. A 10-year retrospective study in 4 pediatric cancer centers in Australia reported L. prolificans as the second most frequent mold infection in children with cancer (27.1% of non-Aspergillus molds), while mucormycosis was uncommon (0.5%) [9]. A retrospective multicenter national study (2014–2017) in France identified 144 IFI in 2721 patients with ALL, and, among these, 6 (4.2%) patients had Fusarium spp. [8]. A single-center study in Rio Grande do Sul in southern Brazil reported an incidence of Fusarium spp. IFI of 14.8 and 13.1 per 1000 patients with ALL and AML [15]. An earlier study involving 7 centers in Brazil and 2 in the US found that the incidence of Fusarium IFI varied by center, but incidence in Brazil was not different from that in the US (6.18 cases per 1000 transplants in Brazil vs. 5.89 cases per 1000 transplants in the US).

Clinical Presentation

IMI are often asymptomatic and diagnosed during work-up of prolonged fever without a source in children with cancer who are receiving antibiotics without improvement. Typically, diagnosis is aided by imaging studies, followed by directed blood and/or body fluid testing, tissue biopsies, cultures, and, increasingly, molecular testing. At times. imaging studies are dictated by new clinical findings, including chest pain, rhino-orbital cellulitis, oropharyngeal ulcerative or necrotic lesions, seizures, changes in mental status, abdominal and/or perirectal pain, swollen lymph nodes, bone/joint pain or swelling, or skin rashes and subcutaneous nodules. Careful and comprehensive daily interrogation and physical exam are of cardinal importance in the evaluation of febrile patients with chemotherapy-induced neutropenia or PIDs. Patients at highest risk, such as A-HSCT patients, receive periodic blood testing even if afebrile: Aspergillus galactomannan antigen (Asp-Gal) and in some centers, 1,3-ß-D glucan. Fungal blood cultures are not typically necessary, since most molds are not recovered from blood cultures except for Fusarium spp. Certain populations at risk may present with specific clinical findings, such as ELBW infants with skin rashes or gastrointestinal pathology.

Aspergillus. Invasive pulmonary aspergillosis (IPA) is the most common clinical syndrome associated with Aspergillus infections. In a recent international cohort of pediatric patients with mold infections, Wattier et al. found 84% of patients with Aspergillus infections had IPA [5]. Earlier, a multicenter retrospective study of pediatric patients with aspergillosis, conducted at 6 large pediatric cancer centers, reported 59% of patients with aspergillosis had IPA. Cutaneous and sinus involvement were reported in 10% and 9.6%, respectively, and 5.9% had CNS involvement. Among 188 patients with sites of infection reported, 39 (20.7%) had ≥ 2 sites involved [2]. Pneumonia is the most frequent infectious syndrome in patients with chronic granulomatous disease (CGD). Aspergillus spp., particularly A nidulans, is commonly encountered [16]. Primary cutaneous aspergillosis, at times with dissemination and poor outcomes, has increasingly been reported in ELBW infants. Gastrointestinal infection is occasionally identified in newborns [10, 11, 17]. Overall case-fatality rates of IA in children with cancer and those receiving a transplant ranges between 20 and 50% but is highly determined by the extent of invasive disease and the severity of immunosuppression [18].

Mucormycosis. Clinical syndromes and patients at risk for mucormycosis are more diverse. Neutropenia and malignancy are common, but type I diabetes, particularly ketoacidosis (10%), and prematurity (17%) are common risk factors [12, 13]. Clinical syndrome varies by age and underlying risk factor. Disseminated disease is common (32%) and the frequently affected sites are cutaneous (27%), gastrointestinal (21%), rhino-cerebral (18%), and pulmonary (16%). Pulmonary infection occurs often in neutropenic patients undergoing induction chemotherapy and in A-HSCT recipients, particularly if A-GVHD is present. Rhino-cerebral is the most common presentation in patients with diabetes, but is also seen in patients with malignancies and A-HSCT. Cutaneous mucormycosis infection results from direct inoculation of fungal spores in devitalized skin and can lead to dissemination. Progression may be slow or fulminant. The presence of a necrotic eschar with surrounding erythema and induration is typical. Gastrointestinal mucormycosis is infrequently diagnosed in living patients (25%). Even then, mortality is high (85%) [13]. Mortality rate varied with the site of infection and host: 96% of patients with disseminated infections, 85% with gastrointestinal infections, and 76% with pulmonary infections died [19].

Other molds. For Fusarium spp., the main routes of infection are inhalation of airborne microconidia or direct inoculation through traumatic injury. Fusarium infections in immune-compromised hosts often affect skin (necrotic papules or nodules), lungs, and sinuses. Hematogenous dissemination is common, with positive blood cultures in 70% of cases. L. prolificans is ubiquitously found as a soil saprophyte, predominantly in arid climates of Australia, southwestern USA, and Spain. Bone/joints, lungs, CNS, and blood were the most frequent affected sites (27%, 27%, 20%, and 18%, respectively). [9, 20,21,22,23]

Diagnosis

Timely diagnosis of IMI is imperative to improve outcomes, and requires a high index of suspicion based on a comprehensive knowledge of risk factors associated with these infections. Clinical differentiation between Mucorales and Aspergillus infections is difficult, yet of significant importance, as treatment for both entities is increasingly divergent with the increased use of newer triazole agents.

Imaging. Unlike adults in whom computerized tomography (CT) findings are fairly specific of mold infection and, at times, suggest differences between Aspergillus (halo sign, air crescent sign) and Mucorales (reverse halo sign, lobar consolidation with cavitation) infections, imaging findings in children are less specific [2, 3, 12]. Still, high-resolution CT (HRCT) chest (with contrast when it can be used safely as contrast enhancement of nodules is suggestive of infection) is strongly recommended in children at high risk of IFI who remain febrile and neutropenic beyond 96 h, even in the absence of focal findings, but certainly in those who develop focal clinical findings, or who have positive screening blood tests. Considering that children often lack the typical radiographic findings observed in adults, finding non-characteristic pulmonary infiltrates or nodules should suggest the possibility of pulmonary IFI and prompt further diagnostic work-up and timely initiation of mold-active antifungal agents (Fig. 1), Imaging of sinuses or abdomen/pelvis in patients without findings indicating pathology at these sites is not routinely recommended. CNS involvement in pulmonary mold disease is frequent in both pediatric and adult patients. Conversely, one-third of patients with cancer or A-HSCT recipients with CNS IMI had no neurological signs or symptoms [24]. Appropriate cranial imaging should be considered in patients with likely pulmonary IMI, even in the absence of neurological features [25].

Fig. 1
figure 1

Imaging studies of patients with IFI. A a micronodule, Aspergillus galactomannan was negative in serum and bronchoalveolar lavage (BAL), next generation sequencing (Karius) showed A. fumigatus (37 DNA copies). B a macronodule, Aspergillus galactomannan in serum was negative, BAL Aspergillus galactomannan and culture were negative, PCR in tissue was positive for A fumigatus. C Four months later, while on voriconazole treatment, the same patient developed a new infiltrate; next-generation sequencing, and later culture revealed Rhizopus delemar. D Serum and BAL were negative for culture and antigen testing; next-generation sequencing identified Rhizomucor miehei. E, F Images of a patient AML who had positive 1–3-ß-D-glucan in serum, negative BAL cultures, and positive blood culture and next-generation sequencing for Fusarium oxysporum

Full size image

Culture and pathology. Obtaining specimens for culture should always be considered, not only for appropriate confirmation of infection but also for identification of pathogen and antifungal susceptibility testing [25, 26]. Bronchoscopy with bronchoalveolar lavage (BAL) should be considered to collect specimens for culture, microscopy, and molecular testing (see below) when HRCT chest shows suspicious findings. Bronchoscopy is difficult to perform in children and requires a pediatric pulmonologist with appropriate training. Surgical intervention has the double value of providing tissue for microscopic diagnosis as well as debulking the infection, and is associated with improved outcomes, as discussed above. Unlike Aspergillus, which has Asp-gal as a biomarker, a diagnosis of mucormycosis is currently limited to biopsy for histopathology, culture, or molecular diagnosis. On microscopy, Aspergillus spp. shows a thin septated hyphae of uniform width with dichotomous filaments branching at acute (< 45°) angles, while Mucorales shows nonpigmented, generally wider, ribbon-like hyphae that are non-septate or pauciseptate, and branching at 90° angles [27]. Tissue from biopsy and BAL should be processed for periodic acid Schiff or Gomori methenamine silver (GMS) staining for microscopy (Fig. 2); specimens should always be cultured in appropriate media. Molecular diagnosis, specifically PCR, will be discussed below.

Fig. 2
figure 2

Histology of biopsies show A Aspergillus spp. in tissue (GMS stain) with thin septated hyphae of uniform width with dichotomous filaments branching at acute (< 45°) angles and B in culture. C Mucor spp. in tissue (GMS stain) and D highlights Mucor's characteristic angio-invasive process

Full size image

Aspergillus galactomannan antigen testing. Asp-Gal is a cell wall component released by all Aspergillus spp. It can be detected with high specificity (galactomannan is not found on other fungi) but only moderate sensitivity by an enzyme immunoassay (Platelia Aspergillus Ag; Bio-Rad, Hercules, CA, USA). Asp-Gal positivity in serum, BAL, and cerebrospinal fluid (CSF) is currently accepted as a mycological criterion for the diagnosis of invasive aspergillosis. Patients with hematologic malignancies or A-HSCT recipients are the best populations for this assay; it does not work as well in SOT. False-negative results in PIDs, specifically in CGD, is common [25, 28]. Currently, patients at high risk for IFI frequently receive mold-active prophylaxis; in this setting, the sensitivity and positive predictive value of Asp-Gal are low, owing to the low pretest probability for the outcome of IFI. Thus, prospective monitoring of patients at risk and receiving mold-active prophylaxis is discouraged. Conversely, Asp-Gal is strongly recommended for diagnosis of aspergillosis in children with clinical signs and symptoms or radiological features on HRCT suggestive of IMI. Wide ranges of sensitivity and specificity are reported, so a negative test does not rule out aspergillosis, and a positive test must be considered with the remaining evidence to avoid unnecessary and potentially toxic treatment [23, 29]. Asp-Gal in BAL with a higher threshold (OD ≥ 1.0) is proposed as an adjuvant tool to routine microbiological studies for diagnosis of IPA. Asp-Gal in CSF (same threshold) has also been used to evaluate for CNS involvement. Concomitant use of piperacillin-tazobactam, amoxicillin-clavulanate, and intravenous immunoglobulin (IVIG) are recognized as potential causes for false positive results.

1–3-ß-D-glucan. 1–3-ß-D-glucan (BG), as a biomarker, allows the detection of a wider array of pathogenic fungi, including Aspergillus spp., Candida spp., Fusarium spp., Trichosporon spp., and Pneumocystis jirovecii, since BG is present in the cell wall of many fungi. Thus, in contrast to Asp-Gal, it lacks specificity. Mucorales do not have BG in their cell wall, so they are not detected by this test. Additionally, BG may be detected in healthy individuals and in certain bacterial infections. Severe mucositis and infusions of platelets, albumin, and IVIG may result in positive results, limiting the utility of the test. Due to limited pediatric data, and poor performance of the test in high-risk patients [low predictive value of a positive test (17–49%) in pediatric studies], BG testing is not recommended in pediatrics as a routine diagnostic test [25, 29, 30].

PCR. Based on data obtained primarily from studies carried out in adults with hematological malignancies, PCR-based diagnostic methods in blood, serum, plasma, or BAL have recently been included in the European Organization for Research and Treatment of Cancer and the Mycoses Study Group Education and Research Consortium criteria as a robust diagnostic test for screening and confirming the diagnosis of invasive aspergillosis in adults.

A systematic review of pediatric studies using blood samples in pediatric cancer patients for screening (3 studies) or diagnosis in children with high suspicion of invasive aspergillosis (8 studies) found broad variability in both settings with specificity (43–85%), sensitivity (11–80%), positive predictive value (PPV) (22–50%) and negative predictive value (NPV) (60–96%) for screening group and sensitivity (36–83%), specificity (0–100%) PPV (0–71%) and NPV (88–100%) in the diagnosis group [31]. Due to these results, a recent international expert group gave a strong recommendation with a moderate quality of evidence for not using fungal PCR testing in blood [32]. A similar recommendation was made in the 2020 guidelines for diagnosis, prevention, and treatment of IFI in pediatric patients for monitoring for IFI in patients at risk, but provided moderate support for the use of PCR in plasma, serum, and whole blood for diagnostic purposes [25]. Simultaneous utilization with Asp-Gal in serum for screening in high-risk patients has increased the specificity of PCR, particularly when PCR is positive in two consecutive tests, similarly for confirmation of diagnosis using PCR and Asp-Gal in BAL [33,34,35].

Quantitative PCR (qPCR) detection of Mucorales DNA in serum, plasma, and BAL has been shown to be a sensitive and early tool for diagnosing mucormycosis. The high sensitivity of qPCR to identify Mucorales is likely related to the large amount of circulating fungal DNA noted in patients with mucormycosis (10–100 times higher than in aspergillosis). Recently, the Fungal PCR Initiative carried out a multi-laboratory evaluation of the performance of different PCR platforms and found good reproducibility and performance, and support for the use of qPCR testing of serum as part of the diagnostic strategy for mucormycosis [37]. Several multiplex methods have been developed for the identification of fungal DNA and are currently being evaluated [35, 38]. Currently, the data are insufficient to make recommendations for PCR in blood or BAL for the diagnosis of mucormycosis, except for PCR methods in the identification of Mucorales in frozen tissue and when molds are seen in formalin-fixed paraffin-embedded tissue [36]. A recent publication reports the use of PCR for the diagnosis of mucormycosis in 5 of 10 patients with hematologic malignancies, 3 of whom survived. The authors speculated that their better survival when compared to a 9-case series from the literature review could be due to earlier, PCR-based diagnosis and initiation of treatment [39].

Next-Generation Sequencing. Microbial cell-free DNA next-generation sequencing (mcf-DNA NGS) is currently being evaluated for diagnosis of IMI. It has shown greater promise for non-Aspergillus IMI, as the sensitivity for proven/probable IPA was low (34%), but it identified a second mold in a few cases. Most patients who were tested by mcf-DNA NGS were receiving Aspergillus-active antifungal agents at the time. When used with Asp-Gal, the combination of these two tests showed a sensitivity of close to 85% to identify IMI [40]. More data are needed before making broad recommendations.

Prophylaxis

Patients receiving intense myelosuppressive regimens for treatment of hematological malignancies or as induction chemotherapy for A-HSCT are at highest risk for IFI, specifically IMI. Systemic antifungal prophylaxis has proven effective in reducing such infections in these patients. There exists a broad and strong consensus that pediatric patients receiving chemotherapy for AML should receive systemic antifungal prophylaxis. For acute lymphocytic leukemia, the recommendation is weaker, since the risk for IFI is treatment protocol- and phase-specific and depends on several factors, such as remission status and corticosteroid utilization, among others. Antifungal prophylaxis hence may be considered for certain subgroups of ALL patients during times of neutropenia. Similarly, antifungal prophylaxis should be administered to pediatric patients undergoing A-HSCT during pre-engraftment/early post-engraftment periods and for those being treated for GVHD. Increasingly, the data support the use of a mold-active agent (echinocandin or mold-active triazole) if prophylaxis is used. Caution is necessary when using triazole agents in children for prophylaxis or treatment, since great interpatient and intrapatient pharmacokinetic variability exists [41,42,43,44,45].

Fluconazole. Fluconazole prophylaxis showed decreased fungal infections (eminently candidiasis) and improved overall survival and decreased fungal-associated mortality in A-HSCT recipients [46, 47].

Voriconazole. A prospective, multicenter, randomized, double-blind trial, which included children ≥ 2 years, compared fluconazole (295 patients) to voriconazole (305 patients) for 100 days (180 for those at highest risk) in A-HSCT recipients and showed no significant difference in IFI (fluconazole 11.2% and voriconazole 7.3% at 180 days, and fluconazole 13.7% and voriconazole 12.7% at 12 months), nor in fungal-free survival (FFS) at 180 days (75% and 78% for fluconazole and voriconazole, respectively at 180 days and 65% and 64% at 12 months). At 180 days, there was a trend to fewer Aspergillus spp. infections in the voriconazole arm (17 vs. 9; P = 0.09). Monitoring for fungal infection was intense. The authors concluded that in the setting of intense monitoring and structured empiric therapy, 6 months FFS and overall survival were not different in A-HSCT receiving fluconazole or voriconazole prophylaxis [48]. Other studies with methodological flaws (single-center, small sample size, retrospective, different indications, etc.) exist, evaluating voriconazole prophylaxis in cancer patients. One such study was a single-center, pre-/post-voriconazole implementation in pediatric patients at high risk for IFIs (treatment of AML). This study found no difference in the incidence of IMI between treatment groups: Aspergillus spp. was the most common IMI when voriconazole prophylaxis was not used, while dematiaceous fungi were most common among children receiving voriconazole, with no cases of invasive aspergillosis identified [49].

Posaconazole. The two largest prospective, randomized, double-blind studies of posaconazole for antifungal prophylaxis included patients > 13 years old, but very few children were enrolled [50]. Similarly, a study comparing posaconazole to fluconazole in patients with severe GVHD enrolled only one patient of 13–18 years old [51]. Both studies used posaconazole suspension, which is no longer in use. The former study demonstrated that posaconazole prevented invasive fungal infections more effectively than comparator agents, and showed improved survival advantage, the latter study showing that posaconazole was similar to fluconazole in preventing fungal infections in patients with GVHD and superiority in preventing aspergillosis. The data from these studies cannot be extrapolated to pediatric patients (low number of subjects 13–18 years and absence of younger children), and the suspension used in this study showed a very low probability of pharmacokinetic/pharmacodynamic (PK-PD) target attainment in pediatric patients [41].

Isavuconazole. No prospective, randomized controlled trial on primary antifungal prophylaxis have been published to date.

Micafungin. A prospective, randomized controlled trial of micafungin (50 mg q 24 h; 1 mg/kg for pediatric patients) compared to fluconazole (400 mg; 8 mg/kg q 24 h) for prophylaxis of IFI during neutropenia in A-HSCT recipients enrolled 882 subjects, approximately 10% of whom were children 6 months–16 years old. The proportion of patients on micafungin who reached primary outcome (absence of fungal infection) was significantly higher for micafungin compared to fluconazole. Importantly, there was one case of probable aspergillosis in the micafungin group compared to 7 (4 proven, 3 probable) in the fluconazole arm [52].

Caspofungin. A multicenter, randomized, open-label study comparing caspofungin to triazole (100 patients on fluconazole and 46 patients on voriconazole) prophylaxis for A-HSCT failed to show a difference between caspofungin and the triazole comparator group, or with each comparator agent individually. The study prematurely terminated enrollment due to futility [53]. A separate prospective, randomized, open-label study comparing caspofungin (253) to fluconazole (255) in children, adolescents, and young adults with newly diagnosed AML also terminated enrollment due to perceived futility (72% of planned enrollment), yet, outcome analysis on patients enrolled at the time of closure showed an incidence of IFI of 3.1% in the caspofungin arm compared to 7.2% in the fluconazole group (P = 0.03) [54].

Treatment

The most common classes of antifungal agents used in the treatment of mold infections include the polyenes (amphotericin B products), azoles (itraconazole, voriconazole, Posaconazole, and isavuconazole), and the echinocandins (micafungin, caspofungin, and anidulafungin). The allylamine antifungal (terbinafine) is sometimes recommended in combination with other agents. Polyenes exert their fungicidal activity by binding ergosterol in the cell membrane, resulting in increased permeability and leakage of intracellular material and cell death. Similarly, azoles target ergosterol inhibiting its synthesis, which is accomplished by inhibiting the lanosterol 14α-demethylase enzyme. Allylamines like terbinafine also interfere with ergosterol synthesis by inhibition of squalene epoxidase. The echinocandins exhibit a completely different mechanism of action as they inhibit cell wall synthesis by inhibiting 1,3-β-D-glucan synthase activity [55].

Treatment recommendations for pediatric patients are based on efficacy in phase II and III trials in adults, the availability of pediatric pharmacokinetic data, safety data, and supportive efficacy data. Currently, treatment of IFI typically starts as empiric or preemptive treatment. A recent multicenter prospective RCT study in Santiago, Chile, showed no difference among 149 subjects in overall mortality, or in IFI between empiric or preemptive therapy after 4 days of persistent fever in high-risk febrile neutropenic patients; the investigators showed a significant lower use of antifungal agents in the preemptive cohort. Most patients (85%) who developed IMI in a recent study had received antifungal agents during the month prior to diagnosis [22].

There exist long-standing accepted basic principles for the management of IMI in adults and children, yet achieving successful treatment can be challenging. These basic principles include:

  • Urgent surgical intervention. Surgical intervention was associated with improved survival in early studies of children with invasive aspergillosis and most importantly in those with mucormycosis [2, 12]. Surgical debridement decreases fungal burden and may reverse low pH and oxygen tension resulting from impaired perfusion and tissue necrosis. Surgery allows for collection of specimens for tissue examination, cultures, and antifungal susceptibility testing [26, 56].

  • Reversal of immunosuppression. Consistently, and whenever possible, reversal or modulation of immune suppression is strongly recommended. This may be difficult, as most patients with IMI have high-risk or refractory (primarily hematological) malignancies or are A-HSCT or SOT recipients. Augmentation of the immune system with growth factors, gamma interferon, or granulocyte transfusions has also been proposed, and should be considered, but has not been systematically studied [25].

  • Prompt initiation of antifungal treatment

  • Therapeutic drug monitoring (TDM). This is increasingly being accepted and more important in pediatrics

Aspergillus

Immediate antifungal therapy is recommended when invasive aspergillosis is suspected in patients at high risk for infection, typically due to prolonged fever (≥ 96 h), or when new fever develops in patients already receiving broad spectrum antibiotics (empiric) or positive galactomannan test/imaging study (preemptive). These patients typically have hematological malignancies or are A-HSCT recipients during pre-engraftment and early post-engraftment phases, or when affected by GVHD, specifically if receiving steroids. Choice of antifungal agents may be impacted by previous or ongoing use of prophylactic antifungal agents. Other high-risk clinical situations exist when invasive aspergillosis is suspected, such as CGD and other PIDs (pneumonia) or extreme prematurity (skin lesions), and treatment should be initiated promptly. Amphotericin B deoxycholate is no longer recommended due to less efficacy and increased toxicity, except in very specific circumstances like renal/bladder infections.

Triazoles

Voriconazole. Following an unblinded RCT showing superiority of voriconazole to amphotericin B deoxycholate for the treatment of invasive aspergillosis [57], voriconazole has been accepted as first-line therapy for invasive aspergillosis, and is recommended in several adult and pediatric treatment guidelines [18, 25, 58, 59]. In January 2019, the FDA extended the indication of voriconazole to include children > 2 years. Voriconazole PK is different and highly variable in children who exhibit linear PK compared to adults, and dosing is difficult (see below), while oral bioavailability is lower (45%). Voriconazole is primarily cleared via hepatic cytochrome P450, with CYP2C19 being the predominant enzymes. Allelic polymorphisms of CYP2C19 likely account for the wide variations in voriconazole clearance and exposure. It has profound interactions with many agents, including some frequently used in transplant (cyclosporine, tacrolimus, etc.); drug levels should be monitored [60]. TDM is strongly recommended: a 2-compartment Michaelis–Menten PK model best described the variability in voriconazole exposure in a study involving 46 children, age 0.8–25 years. The authors further reported that trough serum voriconazole level < 1 mcg/mL was associated with a 2.6-fold increase in odds of death (95% CI, 1.6–4.8; P = 0.020) [61]. Troke et al., reported the relationship between voriconazole PK, minimum inhibitory concentration (MIC), and outcome against Aspergillus spp. The authors concluded that higher free Cavg (AUC/dosing interval)/MIC ratios were associated with a progressively higher probability of response, and that Cmin (trough)/MIC ratio of 2–5 had the strongest correlation with clinical success [62]. Aspergillosis in ELBW newborns presents a challenge unique to pediatrics, as no large body of data exists. The treatment principles remain the same with surgical debridement and rapid initiation of antifungal therapy. A few reports are emerging on the use of voriconazole for these newborns. Consultation with infectious diseases specialists is strongly recommended [10, 11, 17] (Table 1).

Table 1 Pediatric voriconazole dosing for treatment and prophylaxis of invasive aspergillosis
Full size table

Posaconazole. Most data on posaconazole are from studies using the oral suspension which is no longer recommended due to erratic PK, particularly in children [41]. PK data of an IV formulation and an oral delayed-release tablet for patients > 18 years old and > 34 kg have been conducted, and provided dosing information used in more recent trials [63, 64]. Groll et al. also evaluated a powder for oral suspension (PFS) in children 2–17 years [65]. A recent international prospective, randomized, double-blind, double-dummy study, which included children > 13 years, > 40 kg, (only 5 subjects were < 18 years old), compared 300 mg IV or oral delayed-release tablets to IV/oral voriconazole and met non-inferiority criteria [66]. These data resulted in FDA approval in 2021 for posaconazole as a first-line agent in the treatment of invasive aspergillosis for patients > 13 years old. Multiple single-center studies have attempted to evaluate an optimal IV posaconazole dose in pediatric patients [44, 67]. A consensus does not exist, as PFS elicits lower levels than the same IV dose in mg/kg; posaconazole doses ranging from 6 to 10 mg/kg have been proposed, although dosing by body surface area of 225 mg/m2 may provide better exposure across age groups. TDM is strongly recommended in children. Multiple studies have shown a strong exposure–response relationship, and several experts recommend trough levels > 0.7 mcg/ml for prophylaxis, > 1.0 for treatment, and > 1.25 for salvage therapy. A practical approach to posaconazole treatment is to target the highest attainable concentration, as there appears to be a minimal relationship between concentration and toxicity to date [68]. An open-label clinical trial of posaconazole IV/oral delayed-release tablets or powder for oral suspension for the treatment of aspergillosis in children is in progress (NCT04218851). Original oral suspension should not be used due to low probability of target attainment [41]. It must be highlighted that parenteral and PFS formulations may not be available in some countries.

Isavuconazole. Isavuconazole, administered as the prodrug isavuconazonium sulfate, is a newer triazole agent, but currently not approved for use in children < 18 years. Similar to the triazole agents described above, use in children is based on adult clinical information and pediatric PK-PD data. A double-blind RCT comparing isavuconazole 200 mg IV/oral to voriconazole for the primary treatment of mold infections, including Aspergillus spp and other filamentous fungi, showed that isavuconazole was non-inferior to voriconazole but was better tolerated with fewer adverse events. Arrieta et al. published the only report of safety and PK data in pediatrics for isavuconazole [69]. Following a standard 10 mg/kg dose (max 372 mg) IV or oral (≥ 6 years old) of isavuconazonium sulfate (equivalent to 5.38 mg/kg, max 200 mg, isavuconazole), the mean Cmax obtained was comparable for patients 1–11 years at 6.8 mg/l and 6.0 mg/l, respectively, suggesting high bioavailability similar to that of adults. However, adolescents 12–17 years old experienced lower Cmax and daily exposure at the same dosage (5.0 mg/l and 76.8 mg h/l, respectively). Using a pre-specified probability of target attainment (AUC 60–233 mg h/l), approximately 18–21.6% of adolescent had exposure below 60 mg h/l, which is higher than children in the younger age group, 11.4% and 7.2% for children aged 1 to < 6 years and 6 to < 12 years of age, respectively [69]. Currently, there is no clear exposure–response relationship identified and TDM is not indicated, but, in children and adolescents, some experts recommend TDM targeting trough concentrations of 2–4 mg/L. A clinical trial of isavuconazole for the treatment of aspergillosis or invasive mucormycosis in children (NCT03816176) has recently been completed.

Azole-resistant Aspergillus fumigatus complex. The successful outcome of immune-compromised patients with invasive aspergillosis has been improved by the availability of azole antifungal drugs. Furthermore, these are the only oral anti-Aspergillus agents, hence are often used for long-term management of patients with chronic Aspergillus infections (cystic fibrosis, aspergillomas, etc.). The emergence of azole-resistant Aspergillus, particularly the A. fumigatus complex and A. fumigatus sensu strictu, has been linked to two different means for resistance development have been identified: through long-term azole patient therapy and by utilization of environmental fungicidals containing azole agents, mainly in agriculture. The most common mechanism of resistance are point mutations in the Cyp51A gene, which is the target for the azole agents, yet there likely are other mechanisms present since isolates with phenotypic resistance having no Cyp51A mutations are often identified. Environmental resistance appears to be more common in Europe. Mutation TR34/L98H has been identified first in the Netherlands and later in other countries in Europe. TR46/Y121F/289A may have been first found in the US in 2008. A recent surveillance study in Europe (22 centers/12 countries) found an overall prevalence of azole resistance of 3.2% and an incidence of 5.2% [70]. Importantly, mutations in Cyp51A are not associated with a decrease in fitness, and patients infected with azole-resistant A fumigatus have a high mortality rate near 70–80%. Identification of resistance may be delayed, since Aspergillus is often not identified in culture, and, when it is, susceptibility testing is not commonly performed [71]. Recently, a PCR has been developed that includes markers for the detection of TR34/L98H and TR46/Y121F/289A for performance in BAL [72]. Cyp51A mutations reduce activity of all azoles, and, when azole resistance is suspected, L-AmB should be considered, some experts recommend the addition of echinocandin. In areas known to have a high prevalence of environmental Aspergillus resistance (> 10%), treatment should be started as described above. Olorofim, one of several new antifungal agents currently in development, retains excellent activity against azole-resistant Aspergillus [73].

Polyenes

Among the polyene agents, the strongest recommendations for first-line treatment in several treatment guidelines are for liposomal amphotericin B (L-AmB). The strength of this recommendation is less than for voriconazole because available data are those of a high-/low-dose treatment trial which demonstrated no additional benefit of a 10-mg/kg regimen compared to a 3-mg/kg dose, rather than a comparative trial [74]. Better survival with both regimens for possible versus probable/proven aspergillosis suggests earlier treatment is beneficial, but it is possible that some of the patients with positive (halo sign on HRCT) may not have had IPA; this is further complicated in pediatric patients in whom the halo sign is infrequent. TDM is not indicated for L-AmB or other amphotericin B formulations.

Echinocandins

Current recommendations do not recommend use of echinocandins as primary therapy for aspergillosis. Caspofungin and micafungin can be used in settings where azole and polyene agents are contraindicated [25].

Combination therapy

Even though current guidelines do not give a strong recommendation for combination therapy for aspergillosis, they suggest consideration of a combination of voriconazole and echinocandin in the setting of severe disease, particularly in patients with hematological malignancy and those with profound and prolonged neutropenia. A recent study comparing voriconazole monotherapy to voriconazole plus anidulafungin may have suggested some advantage for combination therapy, particularly in patients with probable disease (positive Aspergillus galactomannan) [75]. A recent study in pediatric patients reported that combination therapy was used in 54% of patients with aspergillosis, more frequently triazole and echinocandin (32%) and triazole and polyene (28%) [5].

Mucormycosis

The mortality rate for mucormycosis is quite high, ranging from 40 to 80% [26]. Interpretation of data on treatment of mucormycosis is difficult, since outcomes can be affected by the immune status of the patient, the site of infection, fungal species, antifungal susceptibility, age of the patient, and variability in surgical intervention. A case series suggested that cerebral, gastrointestinal, and disseminated mucor had the highest mortality rates [12]. As discussed above, surgical intervention and immune restoration when possible are important adjunctive treatment of mucormycosis. Surgery improved survival from 50 to 70% when combined with antifungal therapy [76]. Repeated resections or debridement, if necessary for source control, should be pursued whenever possible. Iron chelation with deferoxamine is contra-indicated, but iron chelation with other agents like deferasirox has been explored. In a small, randomized study, patients were allocated to receive liposomal amphotericin B with (11 patients) or without (9 patients) deferasirox; the former had the higher mortality at 30 and 90 days [77]. Population imbalances made interpretation difficult but, currently, deferasirox is not routinely recommended [25]. Because of the rapidity of the growth of Mucor, treatment should be initiated immediately. Antifungal treatment recommendations are based on adult studies and pharmacokinetic pediatric data.

Polyenes

Liposomal amphotericin B (L-AmB). This is considered a first-line agent, with dosing ranges from 5 to 10 mg/kg/day. Most experts would start at 10 mg/kg or escalate rapidly. This recommendation is not based on RCT data [25,26,27]. A few small single-arm studies have evaluated high-dose (10 mg/kg) L-AmB for treatment of mucormycosis [78, 79]. A safety, tolerance, and pharmacokinetic study of high-dose L-AmB in patients with mold infections showed linear PK of amphotericin B for doses up to 10 mg/kg, with no increase in peak concentration (Cmax) at doses of 12.5 and 15 mg/kg. Of note, no significant dose-dependent differences in creatinine have been noted; however, there were more cases of severe hypokalemia among those receiving > 10 mg/kg. Doses < 5 mg/kg are not recommended, and experts recommend against the use of amphotericin B deoxycholate [25, 26]. An important limitation to lipid formulations of amphotericin B is the difficulty in treating fungal infections localized to the kidneys or bladder. A dose of 2 mg/kg L-AmB had tenfold less urinary excretion than 0.6 mg/kg amphotericin B deoxycholate. In fungal renal infections, amphotericin B deoxycholate should be considered instead of lipid formulations. Outside of this clinical situation, amphotericin B deoxycholate should only be used in settings where other treatment options are not available.

Triazoles

These classes of agents offer improved safety, and, where available, the possibility of oral treatment may be attractive when possible susceptibility testing is encouraged.

Isavuconazole. This agent was evaluated in a single-arm, open-label study in patients > 18 years old with invasive fungal infection, including mucormycosis. The patients were matched with controls from a large registry of fungal infections who received amphotericin B products; day 42 mortality in isavuconazole patients (33%) was similar to matched controls (39%) [80]. As a result, isavuconazole was approved in the USA and Europe for first-line treatment of mucormycosis. Recent PK studies in pediatric patients with the IV and oral formulations with isavuconazole have been completed and was discussed above in the "Aspergillus" section  [69]. A study evaluating isavuconazole for the treatment of invasive aspergillosis and mucormycosis in children has recently been completed (NCT03816176). Isavuconazole has also been recommended for salvage therapy.

Posaconazole. The oral suspension of posaconazole has been evaluated for salvage treatment (primary failure of treatment with or intolerance to L-AmB) of invasive mucormycosis, with reported success rates of 60–79% [52, 81]. This suspension is no longer recommended, as PK studies of delayed-release tablets and parenteral formulations for patients > 18 years old showed adequate exposure (Cavg) in > 90% of patients. Not enough data are currently available to recommend the use of posaconazole in children < 18 years old. Pediatric PK studies on IV and PFS formulations showed exposure comparable to adults, but significant age and weight variability was noted [44, 65, 67]. Recently, Fatemizadeh et al. reported on the successful use of posaconazole suspension on a ELBW infant at a dose of 7 mg/kg q 8 h, achieving a steady-state Cmin of 0.9 mcg/ml for the treatment of cutaneous Rhizopus spp. infection following initial treatment with amphotericin B and micafungin [82].

Combination therapy

Combination antifungal treatment does not have proven benefit, but there is theoretical benefit [26]. Echinocandins alone are considered ineffective against mucormycosis. Their mechanism of action inhibits the production of 1–3-ß-D glucan and destabilizes fungal membranes, but Mucorales incorporate very little, if any, BG. However, the combination of echinocandins with amphotericin B is synergistic. It appears that echinocandins can unmask BG and prime neutrophils against mold hyphae, which explains the synergism [83, 84]. A recent multicenter prospective cohort study in children with probable or proven IMI showed combination therapy being used in 41% of patients diagnosed with mucormycosis; triazole-polyene (24%) and polyene-echinocandin (12%) were the most common antifungal combinations used [5].

Other molds

Rare mold infections are increasing as prophylactic regimens decrease the incidence of Candida spp., Aspergillus spp., and Mucorales infections in immune compromised hosts. No RCTs exist to guide the management of these infections. Fusarium spp. are the most common molds causing severe disseminated infections in these patients. Voriconazole and high-dose L-AmB (3–10 mg/kg/day) alone or typically in combination are commonly used. TDM should be used, and high voriconazole levels are encouraged. Addition of echinocandins is frequent. Terbinafine in combination with voriconazole has been used with success. L. prolificans is frequently resistant to currently available antifungal agents, but voriconazole has the best in vitro activity. A combination with terbinafine has shown improved survival. Salvage therapy with miltefosine has been reported [85]. Olorofim, a new antifungal in development, is a promising agent. Scedosporium spp. is responsible for frequent fungal infections in patients with cystic fibrosis, but in immune-compromised hosts it can cause eumycotic mycetoma, and pulmonary, sinus, and CNS disease. Voriconazole-based first-line treatment is recommended. Resistance to amphotericin B is common. Combination therapy should be considered with amphotericin B, echinocandins, or terbinafine. Isavuconazole and posaconazole may be used instead of voriconazole.

New antifungal agents

Several new antifungal agents are currently in different stages of development, which will likely enhance the current antifungal armamentarium. Rezafungin is a next-generation echinocandin with increased solubility and stability and a significantly increased half-life (133 h), allowing for once weekly dosing. Its use for prophylaxis (Candida, Aspergillus, and Pneumocystis) is being explored. SUBA-itraconazole is a reformulated form of the itraconazole tablet, with significant increased (173%) oral bioavailability compared to traditional oral formulation, with minimal food or acid effect on absorption. It has been approved by the FDA for endemic mycoses (blastomycosis, histoplasmosis), and aspergillosis. Olorofim, discussed briefly above, belongs to a new antifungal class, the orotomides, with a novel mechanism of action. They inhibit dihydroorotate dehydrogenase, a key enzyme in pyrimidine synthesis which impairs fungal synthesis of nucleic acids, cell wall, and phospholipid, affecting cell regulation and protein synthesis. Olorofim has a broad spectrum of activity against molds, with particular activity against Aspergillus spp. It displays activity against L. prolificans and Scedosporium spp. Ibrexafungerp is another first-in-class agent (terpenoid) with a similar mechanism of action to echinocandins. It has fungistatic activity against Aspergillus spp. and may be active against L. proplificans and Paecilomyces variotii. Like echinocandins, it importantly has potent activity against Candida spp.; unlike echinocandins, it offers an oral formulation. Fosmanogepix, a prodrug that is metabolized into its active compound, manogepix, targets the fungal specific enzyme, Gwt1, and is responsible for an early step in glycosylphosphatidylinositol-anchor biosynthesis. It has potent activity against Candida spp. and Cryptococcus neoformans, but also molds like Aspergillus spp., Scedosporium spp., and Fusarium spp. Amphotericin B cochleate is a new oral dosage formulation of amphotericin in early stages of development [55, 86]. There are several other new agents in different stages of development which, at this time, escape the scope of this review.

Conclusions

Improved survival of children, adolescents, and young adults with debilitating conditions has greatly expanded the population of people at risk for IFI, specifically molds. New diagnostic modalities are being developed, like PCR and next-generation sequencing, which will help to identify infections sooner, with the expectation of better outcomes if treated earlier. New safer and more active therapeutic agents are being developed to prevent and treat IFI, and a thorough understanding of their PK-PD properties and TDM during treatment will help us to develop personalized regimens for patients of different ages and biometric composition to continue to improve outcomes. We must remain vigilant as we have learnt that while there remain patients with underlying risk factors, new fungal pathogens will emerge.