Role of liposomal amphotericin B in intensive care unit: an expert opinion paper

Abstract

Introduction

Invasive fungal infections (IFI) are frequent in patients admitted to the intensive care unit (ICU). The use of first-line antifungals like triazoles or echinocandins may be limited by the global spread of multi-drug resistance species, drug–drug interactions, low organ penetration, and some safety concerns in case of multi-organ failure. Liposomal amphotericin B (L-AmB) is a polyene drug with a broad activity against mold and yeast and an acceptable safety profile. To outline the role of L-AmB in the treatment of IFI in critically ill patients, a panel of experts was invited to draw up an expert opinion paper on the appropriate place in therapy of L-AmB in different clinical scenarios of patients admitted to ICU.

Methods

A multidisciplinary group of 16 specialists in infectious disease, microbiology, pharmacology, and intensive care elaborated an expert opinion document through a multi-step approach: (1) the scientific panel defined the items and wrote the statements on the management of IFI in ICU, (2) a survey was submitted to an external panel to express agreement or disagreement on the statements, and (3) the panel reviewed the survey and implemented the final document.

Results

The final document included 35 statements that focused on epidemiology and microbiological rationale of the use of systemic L-AmB in critically ill patients and its potential role in specific clinical scenarios in the ICU.

Conclusion

Systemic L-AmB may represent an appropriate therapeutic choice for IFI in ICU patients with different underlying conditions, especially when the use of first-line agents is undermined. This expert opinion paper may provide a useful guide for clinicians.

Introduction

Invasive fungal infections (IFI) represent a life-threatening condition in patients admitted to the intensive care unit (ICU). The incidence of IFI in critically ill patients is rising, while attributable morbidity and mortality remain high [1, 2]. Reasons may lie in the higher complexity of care of patients with a major risk for IFI, including immunocompromised patients or those with severe medical or surgical comorbidities [3]. Other risk factors may be related to the extensive use of broad-spectrum antibiotics and invasive procedures which may favor tissue invasion by disrupting the integrity of epidermal and mucosal barriers [3, 4].

The epidemiology of IFI in the ICU is changing. Candidiasis is still the most common fungal infection, though a shift towards non-albicans species has been observed [5,6,7]. Moreover, the rate of invasive aspergillosis (IA) in critically ill patients is increasing [8]. Notably, the evidence of association between severe viral respiratory infections and IA warned about the emergence of new categories susceptible to IA without the “classical” risk factors like neutropenia or transplantation [9, 10]. Infections caused by rare molds like Mucorales species, Fusarium species, Scedosporium species, or Lomentospora prolificans are also standing out [11, 12]. In these cases, therapeutic management is challenging due to the lack of rapid diagnostic assays, the limited availability of antifungal susceptibility testing (AST), and poor clinical evidence about the effectiveness of current treatment options [13].

Finally, the worldwide spread of antifungal resistance to first-line agents like fluconazole, triazoles, and echinocandins is of great concern since currently alternative options are limited [14]. Outbreaks of azole-resistant Candida albicans or C. parapsilosis as well as echinocandin-resistant C. parapsilosis or Pichia kudriavzevii (formerly C. krusei) in ICUs are described worldwide [6, 14]. The recent emergence of nosocomial infections by C. auris is of great concern because of its environmental adaptability and multi-drug-resistant profile [15, 16]. Not least, the prevalence of azole-resistant A. fumigatus is increasing around the world also involving ICU patients with IA [17]. Even if many new antifungals are in the pipeline, robust data on their efficacy in critically ill patients are currently limited [18].

Liposomal amphotericin B (AmBisome®, L-AmB) is a polyene agent comprised of conventional amphotericin B included in liposomal unilamellar vesicles. By binding to ergosterol, amphotericin creates pores in the fungal cell membrane, leading to ion leakage and cell killing [19]. L-AmB has a wide spectrum of activity on numerous fungal species including Candida species, Aspergillus species, Cryptococcus, Rhizopus species, and other rare molds [19, 20]. Indeed, systemic L-AmB demonstrated a safer profile compared to conventional amphotericin B formulations, with a lower rate of nephrotoxicity and infusion reactions [21].

Guidelines consider L-AmB a reasonable alternative in case of refractory or resistant candidiasis and aspergillosis, as well as a first choice for mucormycosis and infections by other filamentous fungi [13, 22, 23]. However, the contemporary landscape of invasive mycosis in ICU is revealing tangible limitations in the use of current first-line agents [24].

For these reasons, a committee of specialists skilled in infections in critical care was called to elaborate an expert opinion document to address the use of systemic L-AmB for the most common IFI affecting patients admitted to ICU, focusing on specific clinical settings.

Materials and methods

The scientific panel included 16 specialists in infectious diseases, microbiology, pharmacology, and intensive care selected based on their clinical expertise and scientific publications:

• Infectious diseases: P. Viale (scientific coordinator), M. Bartoletti (scientific secretary), M. Giannella (scientific secretary), M. Bassetti, F.G. De Rosa, M. Falcone, P. Grossi, M. Mikulska, and C. Tascini

• Intensive care: A. Cortegiani, G. De Pascale, M. Girardis, P. Navalesi, and B. Viaggi

• Clinical pharmacology: F. Pea

• Microbiology: M. Sanguinetti

The methodology for statement elaboration and approval was established in October 2023. A multi-step strategy was chosen to formulate an expert opinion document.

During the first meeting, the panel identified the clinical items and the open issues concerning the management of ICU patients at risk of invasive candidiasis and mold infections and the potential role of systemic L-AmB in these settings (Table 1).

Table 1 Initial items on the management of IFI in ICU and the role of systemic LAMB discussed during the first project meeting

Then, the panel members were divided into subgroups based on specific expertise to produce one or more statements for each item or patient setting (Table 1). The formulation of each statement was supported by a narrative review.

The initial statements were finally reviewed by the whole panel until a general agreement was reached.

In the second step, the statements were tested by an external panel of Italian physicians selected based on proven clinical experience and scientific relevance in the field of infections in the ICU. Of 67 clinicians invited, 51 participated in an online survey. The external panel expressed the level of agreement or disagreement with each statement through a 9-point scale, where 0 points corresponded to “strongly disagree” and 9 points to “strongly agree”.

The results of the survey did not aim to change the content of the statements; however, the statements receiving less than 8.0 of the average rate of agreement were discussed by the scientific panel before their inclusion in the final document.

Results

The scientific panel formulated 35 statements on the general management of IFI in the ICU and the role of LAMB. Table 2 details the statements on the use of systemic LAMB. Overall, the statements received a high level of agreement (median rate 8.0) from the external panel. The statements receiving an average score < 8.0 are marked with an asterisk (*) in the text. These statements and their revisions are shown in Table 3.

Table 2 Summary of statements on the use of systemic LAMB in ICU

Table 3 Review of statements receiving an average rating < 8.0 in the external panel survey

General statements about the role of liposomal amphotericin B

• 1. Considering the scientific evidence currently available, making a univocal decision about treatment choice for every IFI is basically impossible. Moreover, patients admitted to ICU may have specific risk factors for IFI as well as severe impairment of one or more organ functions may affect the antifungal treatment. Thus, every decision regarding antifungal drug choice for severe or complicated infections in critically ill patients should be individualized based on the simultaneous evaluation of epidemiological, microbiological, pharmacological, and clinical variables.

• 2. Inside the antifungal armamentarium, L-AmB represents a valuable choice in several different settings and fungal infections, thanks to its wide antifungal spectrum of activity, limited propensity to develop resistance, low impact in terms of drug-drug interactions, good capability of overcoming biological barriers, no need for therapeutic drug monitoring (TDM) and acceptable safety profile.

Microbiology

Background

The role of microbiological biomarkers in diagnosing IFI in the ICU is highly debated. The characteristics, strengths, and limits of the main tests are shown in Table 4.While molecular and antigen-based methods have improved the speed and sensitivity of diagnosing IFI, the classical culture of clinical samples remains important for confirming the diagnosis and for species identification [39]. However, classical cultural techniques may take several days to yield results, which can delay the initiation of target therapy. Moreover, the sensitivity of cultures may be influenced by various factors, including previous exposure to antifungals or accuracy in sample collection [40] (Table 4).Standardized antifungal susceptibility testing (AST) includes different methods performed on positive cultures; however, new resistance molecular tests are getting into clinical practice [25]. Although not ubiquitously available, the use of AST may be critical to guide antifungal therapy in the ICU. Patients admitted to ICU may have an increased risk for resistant infections, due to both individual factors (i.e., immunosuppression, previous antifungal exposure, or long hospitalization) [26, 27] and environmental conditions (i.e., large use of azoles in agriculture, the crisis of ICUs during COVID19 pandemic) [14, 17]. Of note, outbreaks of fluconazole-resistant C. parapsilosis, multi-drug-resistant C. auris, or azole-resistant A. fumigatus have been reported in ICUs worldwide [17, 26, 28].

Table 4 Characteristics of the main microbiological tests for IFI in ICU

Statements

• 3. Epidemiology of IFI is changing due to several factors including the better performance of microbiological diagnosis, the increased numbers and diversity of susceptible patients (i.e., COVID-19, biologics), the exposure to antifungals both in the individual and in the environment, and the changing climate. Emerging infections/resistance patterns mandate the need for timely and accurate diagnostics as well as for species identification and detection of antifungal resistance. In other terms, access to mycology laboratory expertise is key for the proper management of IFI.

• 4. Despite the considerable variability in populations and reference criteria employed, investigations into laboratory assays for diagnosing IPA consistently revealed GM from BALF a superior diagnostic accuracy over serum GM. Additionally, both BALF and serum BDG demonstrated less-than-ideal specificity.

• 5. Quantitative GM testing, especially in BALF, is a valuable and widely used biomarker for the diagnosis of IA in the ICU. However, results should always be interpreted in the context of clinical, radiological, and other laboratory findings.

• 6. Polymerase chain reaction test from BALF can be a valuable tool for the diagnosis of aspergillosis in the ICU, especially in high-risk and immunocompromised patients. However, the sensitivity and specificity of PCR can vary depending on the patient population and the specific PCR method used. Standardization of protocols for DNA extraction and PCR assays is important for improving diagnostic accuracy (i.e., species identification, resistance genotypes…).

• 7. Beta-glucan testing can be a valuable tool for diagnosing IC with or without candidemia in ICU patients. Based on its high NPV, BDG should be included in the decision tree aimed to exclude systemic Candida infection.

• 8. Classical culture plays an important role in the diagnosis of IA by providing a definitive identification of the pathogen and guiding appropriate treatment strategies.*

• 9. Blood cultures are the gold standard for diagnosing candidemia in the ICU. They are mandatory not only for microbiological diagnosis but also for the identification of causative species, testing sensitivity, and monitoring the timing of treatment response.

• 10. Polymerase chain reaction tests, including pan-fungal PCR assays and plasma cell-free DNA fungal PCR panels, can provide sensitive and specific detection of various fungal pathogens beyond Candida and Aspergillus species. These tests have the potential to aid in the early and accurate diagnosis of fungal infections, leading to improved patient outcomes. *

Pharmacology

Background

Liposomal amphotericin B is characterized by a concentration-dependent fungicidal activity [20]. In experimental animal models, the main determinant of efficacy was found to be the maximum concentration (Cmax)/minimum inhibitory concentration (MIC) ratio [19]. Studies evaluating the pharmacokinetic profile of L-AmB in critically ill patients are rather limited [19, 63, 64]. From the available data, there is a certain interindividual variability, but this does not appear to be attributable to any specific pathophysiological condition. No correlation was found with renal function, albuminemia, and/or Sequential Organ Failure Assessment (SOFA) score [63]. The Cmax and the area under curve (AUC) levels achievable in critically ill patients during treatment with doses of L-AmB of 3–5 mg/kg/day are quite like those found in healthy volunteers and/or other patient populations [63]. Furthermore, maximum concentration (Cmax) and AUC do not appear to be influenced by the application of continuous renal replacement therapy (CRRT) [64, 65]. It has been reported some case reports that during Extra-Corporeal Membrane Oxygenation (ECMO) a certain increase in the volume of distribution (Vd) can occur [66,67,68]; however, available PK data on LAmB in ECMO are few and controversial [69]. Some authors suggested that using doses per/kg of total body weight in patients with morbid obesity could cause an increased risk of nephrotoxicity, especially at a dose of 5 mg/kg/day [70,71,72]. Overall, considering these data, it is believed that in critically ill patients, the maximum dose of L-AmB could be 5 mg/kg/day with a ceiling dose of 500 mg in patients weighing > 100 kg.

A recent meta-analysis analyzed 10 single- or double-blind, randomized, controlled, clinical trials that included a total of 1661 patients treated with high doses of L-AmB (> 5 mg/kg/day; range 6–15 mg/kg/day) compared to standard doses of L-AmB (3 mg/kg/day, 4 studies) or of amphotericin B deoxycholate (0.7–1 mg/kg/day, 3 studies) or of posaconazole (200 mg q6h oral suspension, 1 study) or in the absence of antifungals (1 study) [73]. Therapeutic efficacy was evaluated as the primary outcome, while mortality, survival ≥ 10 weeks, and adverse reactions were evaluated as secondary outcomes. The use of doses of L-AmB > 5 mg/kg/day was not associated with an advantage in terms of clinical outcome. In particular, the analysis of the 3 comparative studies concerning the treatment of IA did not demonstrate any statistically significant advantage for the high doses in terms of therapeutic efficacy (OR = 0.35, 95% CI 0.06–2.12, P = 0.25). By contrast, the use of high doses was associated with an increase in mortality, a reduction in long-term survival (≥ 10 weeks, OR = 0.57, CI 95% 0.34–0.94, P = 0.03) and an increase in adverse events (including renal failure). Out of specific indications for using a high-dose single shot or loading dose of L-AmB [74], the only setting in which the hypothesis of using high daily doses of L-AmB > 5 mg/kg continues to be postulated is that of mucormycosis [75]. However, a recent retrospective, multicenter study analyzing 82 confirmed and probable cases of mucormycosis collected between 2015 and 2022 in 51 Japanese hospitals concluded that the use of high doses > 5 mg/kg/day did not improve survival. Conversely, a single 10 mg/kg dose may be considered a good option for treating visceral leishmaniasis [76, 77] and/or cryptococcal meningitis [74].

The risk of nephrotoxicity of amphotericin B deoxycholate is due to the accumulation that occurs in the renal tubular cells with this formulation. By contrast, L-AmB has a much lower Vd than the deoxycholate formulation, and this results in a lower propensity of accumulation and a lower risk of toxicity. This is because the liposome, by acting as a reservoir and by remaining intact until contact with the fungal membrane, retains the amphotericin B in its wall and may prevent its accumulation at the renal level [19]. In a comparative meta-analysis against amphotericin B deoxycholate including 10 studies with a total of 2172 participants, L-AmB was found to be significantly safer than conventional amphotericin B in terms of increase in serum creatinine over twofold the baseline value (RR 0.49, 95% CI from 0.40 to 0.59) [21]. In the specific context of critically ill patients, a recent prospective phase 2 study enrolling 40 adult patients at high risk of intra-abdominal candidiasis (IAC) after major abdominal surgery demonstrated that pre-emptive therapy with a single 5 mg/kg dose of L-AmB, followed by prompt withdrawal in case of negative baseline BDG result, was a safe and effective approach [78]. A retrospective clinical study evaluated the usage and occurrence of adverse reactions during L-AMB therapy in patients undergoing renal replacement therapy (RRT). In total, 24, 19, and 842 cases were included in the hemodialysis (HD), CRRT, and non-RRT groups, respectively. After propensity score matching, the average daily and cumulative dose, treatment duration, and dosing interval for L-AMB were not significantly different and the incidence of adverse events did not markedly differ among the groups [79].

Statements

• 11. The recommended dose of L-AmB for most indications in critically ill septic patients is 3 mg/kg, with a maximum of 5 mg/kg/day (a dose ceiling of 500 mg is recommended in patients weighing > 100 kg). Daily doses of L-AmB > 5 mg/kg are not associated with a significant benefit in terms of clinical outcome in any type of fungal infection and could increase the risk of nephrotoxicity and hypokalemia. However, a single 10 mg/kg dose could be considered for treating visceral leishmaniasis and/or cryptococcal meningitis.

• 12. The risk of nephrotoxicity of L-AmB at a dose of 3–5 mg/kg/day is much lower than that of amphotericin B deoxycholate.

• 13. In critically ill patients with renal dysfunction and/or requiring hemodialysis or continuous renal replacement, no dosing adjustment of L-AmB is necessary due to the fact that its elimination is non-renal and the incidence of adverse events did not markedly differ from non-RRT groups.

Specific clinical settings

Molds and SARS-CoV-2 and/or influenza virus coinfections

Background

Both severe influenza and severe/critical COVID-19 are associated with a higher risk for invasive pulmonary aspergillosis (IPA). These conditions were named influenza-associated pulmonary aspergillosis (IAPA), and COVID-19-associated pulmonary aspergillosis (CAPA), respectively. Complex pathophysiological interactions involving viruses, the damaged lung parenchyma, immune cells, and Aspergillus spp. were demonstrated. The virus-induced injury and the following activation of the immune cells can facilitate the progression from contamination with Aspergillus conidia to tissue invasion and potentially lead to the angio-invasive phase [80, 81]. The ability of the macrophages to destroy Aspergillus conidia seems to be impaired in case of high viral burden [80, 81]. For this reason, IPA associated with respiratory virus is considered a specific entity in critically ill patients, called virus-associated pulmonary aspergillosis (VAPA) [10, 82]. Clinical practice guidelines and guidance documents for the diagnosis and management of both IAPA and CAPA were released [83, 84]. In patients with severe viral pneumonia, respiratory failure, and need for respiratory support, a diagnosis of IPA should be pursued. Galactomannan optical density index (ODI) on BALF or other deep respiratory specimens should be measured in every patient at ICU admission and serially once a week. As for the use of antifungal prophylaxis in this setting, current clinical evidence does not justify this practice [85, 86]; indeed, the incidence of CAPA and IAPA may vary significantly across different geographical areas [87].

Statements

• 14. In patients with severe viral pneumonia, respiratory failure, need for respiratory support and no other risk factors for IPA, initiation of anti-mold treatment should be postponed until microbiological diagnostic criteria have been addressed. On the contrary, in patients with severe viral pneumonia and other risk factors for IPA (e.g., corticosteroid therapy, COPD, immunosuppression) empiric treatment should be considered. *

• 15. Widespread anti-mold prophylaxis in critically ill patients with viral pneumonia is not currently justifiable by available evidence. *

Patients on therapy with corticosteroids or immunomodulatory drugs

Background

The chronic use of high-dose corticosteroids has been defined as a risk factor for pulmonary aspergillosis for decades. Indeed, chronic therapy with steroids is one of the host criteria of the EORTC-MSG and AspICU algorithm for the diagnosis of IPA [88]. More recently, corticosteroid therapy was found as a peculiar risk factor for developing IAPA in patients with severe influenza [9]. Although dexamethasone was demonstrated to reduce mortality in severe/critical COVID-19 patients, its use was associated with a higher risk of developing CAPA in several observational studies [89]. Dexamethasone seems to reduce the macrophages’ ability to prevent A. fumigatus germination, which may be correlated with fast fungal growth, destruction of macrophages, and induction of an anti-inflammatory cytokine profile. Moreover, other drugs associated with reduced mortality in severe/critical COVID-19 patients, such as anti-interleukin (IL)−6 (e.g., tocilizumab) were associated with a higher risk of developing CAPA [90].

Statement

• 16. Chronic therapy with corticosteroids or immunomodulatory drugs should lead to a high index of suspicion of IPA in critically ill patients with pulmonary infiltrates, driving an early diagnostic approach.

Chronic obstructive pulmonary disease

Background

Patients with COPD are recognized as at higher risk of developing IPA. However, in critically ill patients with COPD, respiratory failure, lung consolidations, and positive Aspergillus tests from the respiratory tract (either culture or GM), the discrimination between Aspergillus colonization or infection may be hard. Since IPA prognosis in critically ill patients is quite poor, the use of algorithms including all those findings may foster early diagnosis and appropriate antifungal therapy [88].

Furthermore, emerging evidence suggests that adopting a pre-emptive strategy in critically ill non-neutropenic patients, particularly those with COPD, may result in significant clinical benefit. This pre-emptive approach is based on the early use of microbiological biomarkers (e.g., GM in respiratory samples, Aspergillus PCR, and BDG assay) and consistent lung imaging [91]. More recently, a risk‑predictive model for IPA in patients with acute COPD exacerbation was proposed, which included serum albumin < 30 g/L, GOLD severity classes III–IV, steroid treatment in the previous three months, and broad-spectrum antibiotics for more than 10 days in the last month [92].

Statement

• 17. Patients with COPD are at higher risk of developing IPA. Therefore, a prompt diagnostic approach must be pursued in any case of infection-related respiratory worsening.

Diabetes

Background

Diabetes mellitus is the leading comorbidity in immunocompetent patients with mucormycosis [93]. Considering that the global prevalence (age-standardized) of diabetes rose from 4.7 to 8.5% in the last 50 years, an estimated 500 million adults are living with diabetes today, with the greatest increment in countries with a valuable circulation of Mucorales such as China, Brazil, Japan, Mexico, Egypt, and India [93,94,95]. Rhino-orbital-cerebral mucormycosis is the most frequent presentation among these patients, even in the absence of underlying conditions of immunosuppression [94]. Of note, COVID-19 pneumonia was described as an adjunctive risk factor for mucormycosis in diabetic patients [96]. The first step of management of mucormycosis should be a high clinical and radiological suspicion and prompt performance of both microbiological and histopathological investigations on tissue samples. However, the severity of infection along with the long processing time of diagnostic tests on tissue imposes an early introduction of empirical antifungal therapy [97]. Moreover, early antifungal administration seems not to affect the yield of histopathology or cultures [98]. The first-line agent for any organ involvement should be high-dose L-AmB and slow dose increment should be avoided [97, 99]. However, a recent retrospective study on 82 patients with mucormycosis did not show better survival of patients receiving L-AmB dose > 5 mg/kg/day versus 5 mg/kg/day [100]. The use of isavuconazole or posaconazole is mainly recommended as second-line or salvage therapy [97, 101]. Of note, clinical data on the efficacy of a combination therapy with amphotericin plus azoles or echinocandins are controversial to support this strategy [97, 102, 103]. Surgery is a cornerstone of the treatment and should be performed whenever feasible [104, 105]. Finally, correction of the predisposing factor including achievement of an adequate glycemic control is critical for the containment of the infection [106].

Statements

• 18. In the last 50 years, diabetes has evolved as one of the major risk factors for mucormycosis, while more recently, underlying malignancy, severe immunodepression conditions, and SARS-CoV-2 infection emerged as important risk factors.

• 19. L-AmB demonstrated efficacy in the treatment of mucormycosis with various organ involvement patterns. The daily dose should be 5 mg/kg per day.*

End-stage liver disease

Background

Increasing data are documenting cases of IPA among critically ill patients with acute liver failure or chronic cirrhosis [107]. Susceptibility to IPA may be related to immune dysfunction associated with liver failure, affecting both innate and adaptive immunity, along with the low platelet count, which has a growth-inhibiting effect on Aspergillus species [108]. The real incidence of IPA in patients with acute liver failure is probably underestimated, except for severe alcoholic hepatitis where incidence is about 15% and mortality almost 100% [109,110,111]. The rate of IPA in patients with end-stage liver disease achieved up to 14%, including those with Child–Pugh score C, a high model for end-stage liver disease (MELD) values/liver failure grade and concomitant COPD. Most of them require invasive mechanical ventilation and renal replacement therapy [109]. Interestingly, in a large cohort of cirrhotic patients admitted to the ICU (n = 986), 60 had a positive respiratory culture for Aspergillus spp, with a 28% rate of proven/putative IPA and 71% mortality rate [112]. Indeed, in critically ill patients with liver failure (especially Child C cirrhosis), the presence of compatible clinical signs and a positive GM antigen (ODI ≥ 1) on BALF, may support the diagnosis of probable IPA [88].

The ESCMID-ECMM-ERS guidelines recommended the use of L-AmB for IPA in patients with liver insufficiency [55]. This consideration relies on the possible hepatotoxicity of azole treatment in the presence of liver failure [80].

Statement

• 20. In critically ill patients, acute on chronic liver failure and decompensated cirrhosis are recognized main risk factors for IA.*

Therapeutic approach to mold infections in patients with severe viral pneumonia, chronic corticosteroids or immunomodulatory therapy, COPD, diabetes, and end-stage liver disease

Statements

• 21. Anti-mold therapy with L-AmB could be preferable over azoles in case of treatment failure and could be proposed as the first-line option (i) in geographic areas with a high prevalence of azole resistance (ii) in patients at higher risk for hepatotoxicity (i.e., end-stage liver disease) in subjects taking drugs having clinically relevant drug-drug interactions vs. azoles, (iii) in setting having no possibility of performing voriconazole TDM.*

• 22. The interindividual pharmacokinetic variability of L-AmB in critically ill patients is expected to be limited so that TDM is not needed.

Solid organ transplantation

Background

The incidence of IFI and distribution of pathogens vary according to the type of transplant and local epidemiology [113, 114]. IFI incidence is usually higher after small bowel, lung, and liver transplantation compared with other types of SOT [113,114,115]. Candida species and Aspergillus species are the main pathogens. Overall IC is the prevalent type of IFI after abdominal transplantation, while IA is the main IFI after lung transplantation [113,114,115]. Studies assessing in deep the epidemiology of candidemia/IC in SOT recipients have shown a shift toward non-albicans Candida species over time with an increasing prevalence of N. glabratus and C. parapsilosis [116, 117], species associated with reduced susceptibility to azoles. A. fumigatus sensu strictu is the prevalent cause of IA in SOT recipients, with A. terreus and A. flavus representing less than 20% of isolates [118]. Azole resistance is an emerging issue in IA, mainly after SOT [44]. It has been associated with the isolation of Aspergillus cryptic species or with the selection of azole-resistant-A. fumigatus mediated or not by non-environment associated mutations and linked or not with prolonged exposure to azoles [119, 120]. Usually, IFI occurs within the first 6 months after transplantation; however, delayed episodes are also observed. A complicated post-transplant course is generally associated with early IFI, while persistent profound immunosuppression is the main predisposing factor for late IFI [121]. Specific risk factors for IC and for IA have been described in each type of SOT (i.e., high-MELD and choledocojejunostomy for liver transplantation; single lung and bronchial stent or ischemia for lung transplantation) [122]. A recent metanalysis aimed at identifying risk factors for IFI within the first year after SOT, showed reoperation, post-transplant renal replacement therapy (RRT), and Cytomegalovirus disease as having a high certainty of evidence and strong associations (relative effect estimate ≥ 2) across all types of SOT [123]. Antifungal prophylaxis is the main strategy to prevent IFI after SOT. Old studies assessing the universal prophylaxis showed a reduced incidence of IFI and IFI-associated mortality, but no impact on overall mortality, on the other hand, a shift toward non-albicans Candida species was observed [124]. Thus, a targeted approach is currently recommended limiting the use of antifungal prophylaxis to patients at high risk for IFI [122]. Indeed, this approach has been shown to be effective and feasible in real life [125]. However, the choice of the best antifungal agent for prophylaxis in SOT recipients at high risk of IFI is controversial [126]. One RCT including liver transplant (LT) recipients at high risk for IFI showed no difference between anidulafungin and fluconazole, but it was limited by a low rate of IFI (only 2 episodes of IA in the fluconazole group) [127]. One meta-analysis did not find a difference in preventing IFI between amphotericin B and fluconazole, but it included very old studies [128]. One propensity-matched multicenter cohort study showed no difference in the overall rate of IFI between caspofungin and fluconazole after LT. However, after adjusting for confounders, caspofungin was associated with a lower rate of IA [129]. High-risk patients receiving L-AmB as antifungal prophylaxis after LT showed the lowest risk of breakthrough IFI compared with those receiving no prophylaxis, fluconazole, or echinocandins in a multicenter cohort study [130]. An increased risk of breakthrough IFI associated with echinocandin prophylaxis after LT was also confirmed by a meta-analysis [131]. Finally, considering drug–drug interaction, the need for TDM, and safety issues, triazoles are considered not easy to handle after SOT, mainly in LT recipients. For all the above considerations, the use of pulsed doses of L-AmB is considered the better option mainly in the setting of LT. In a phase II uncontrolled trial including 76 high-risk LT recipients, prophylaxis with L-AmB administered at the dosage of 10 mg/kg once weekly was shown to be safe with only 3 patients developing acute kidney injury unrelated to the study drug; in addition, the IFI rate was significantly lower than that observed in a historical control group (2.6% vs. 11.8%, p = 0.03) [132]. Recommendations about the therapeutic management of IC and IA in SOT recipients are the same as for non-SOT recipients [133]. For IA, isavuconazole has been shown to be safe and effective in the management of SOT recipients with invasive mold infections [134]. Compared with voriconazole and posaconazole, isavuconazole has fewer drug-drug interactions with immunosuppressant drugs. A recent single-center retrospective cohort study including 68 patients (51 lungs, 14 hearts, and 3 heart/lung transplant recipients) investigated the concentration to dosage ratios (C/D) of immunosuppressants when starting isavuconazole de novo or shifting to isavuconazole from other azole treatment. The authors observed a temporary doubling of tacrolimus exposure, as well as a required dose decrease for cyclosporine and sirolimus when starting isavuconazole de novo. Tacrolimus C/D increased by 110% at day 3 in patients started on isavuconazole de novo. When transitioning from other azoles, tacrolimus and cyclosporine required about twice the initial dose [135]. Finally, although routine TDM of isavuconazole exposure is not routinely recommended, in patients with severe liver disease, an increased exposure may occur, thus requiring dosage adjustment [136]. L-AmB is considered the best option in patients in whom first-line therapy is associated with an unacceptable adverse-event profile, drug–drug interaction, or risk for resistant/refractory disease [133, 137].

Statements

• 23. In SOT recipients, a targeted (risk-based) approach to antifungal prophylaxis is recommended. Clinically relevant drug–drug interactions, safety concerns, and rates of breakthrough infections are all issues to be taken into account when choosing an antifungal agent for prophylaxis. In this regard, L-AmB may be considered a suitable option. *

• 24. Drug–drug interactions with immunosuppressive drugs could sometimes represent a relevant issue when treating IFI with azole antifungals after SOT. In this regard, L-AmB could be a valuable alternative option for the empirical treatment of IFI. *

• 25. Regarding IC in SOT recipients, L-AmB could be considered a reasonable alternative to echinocandins. *

• 26. Since antifungal stewardship has emerged as an important component of quality in managing IFI, the application of a targeted prophylaxis or pre-emptive antifungal treatment is a valuable approach in every transplant setting, including lung transplant. *

Hematologic malignancies

Background

Patients with long-term neutropenia following chemotherapy for acute myeloid leukemia (AML) and patients undergoing allogeneic hematopoietic stem cell transplantation (HSCT) are at high risk of contracting IFIs. Moreover, new risk categories are emerging, for example, patients treated with immunotherapy or chimeric antigen receptor (CAR) T cell therapy who may develop prolonged phases of severe neutropenia during and following the treatment [138]. The clinical efficacy of antifungal prophylaxis in high-risk patients has been demonstrated in randomized controlled trials and is now recommended in international guidelines [22, 139,140,141,142]. Although this strategy has resulted in a decline in the incidence of IFIs in high-risk hematology patients, a subset of such patients still develops breakthrough IFIs (bIFIs) [143]. Cohort studies conducted after the introduction of posaconazole as the standard of care for prophylaxis in this setting further highlighted the development of posaconazole-associated bIFIs with variable incidence rates depending on the study (0–10.9%) [144]. IA caused by A. fumigatus is most often represented among these bIFIs, but IA caused by non-fumigatus species and bIFIs caused by non-Aspergillus molds have also been reported including several cases of mucormycosis and fusariosis [12, 145,146,147,148,149,150]. The occurrence of bIFI in this setting may be explained by three clinical scenarios, in addition to a severe immune deficit or increased fungal virulence [143]: (i) sub-therapeutic drug levels in patients receiving azole prophylaxis, (ii) azole-resistant Aspergillus fumigatus, and (iii) intrinsic posaconazole-resistant IFI (some Mucorales strains, Fusarium, or some other rare molds). In these scenarios, the choice of treatment should be individualized according to several factors, but in most cases, the initiation of treatment with L-AmB is appropriate as this drug provides broad-spectrum coverage against azole-susceptible and azole-resistant Aspergillus, various species of Mucorales, Fusarium, some—but not all—other filamentous fungi and common or rare yeasts [13, 23]. The treatment should be continued based on antifungal susceptibility testing results, if available.

Although there are several new antifungal agents in the pipeline, triazoles continue to be the mainstay of therapy for the treatment and prevention of IFIs in hematological patients, but their clinical use is complicated by variable pharmacokinetics and drug–drug interactions. Therefore, there is increased recognition of the need for antifungal stewardship and practical guidance for TDM for patients with IFIs.

Given the marked intra- and inter-patient pharmacokinetic variability of voriconazole and the association of plasma exposure with both efficacy and toxicity, voriconazole concentrations should be routinely monitored in patients receiving this agent for prophylaxis or treatment [151, 152]. As previously reported, even if it was generally accepted that isavuconazole has lower variability in terms of pharmacokinetics, recent studies suggest that, especially in the ICU setting, isavuconazole plasma concentrations may vary in critically ill patients and significantly lower isavuconazole levels were observed in patients with elevated body mass index and higher SOFA score [153,154,155]. Overall, these studies indicate that TDM for azole is strictly necessary in the ICU setting to optimize efficacy and reduce unintended side effects [156]. Therefore, in centers where TDM is not available an alternative treatment to azole such as L-AmB could be considered when treating a critically ill patient with suspected or confirmed invasive mold infection.

Considering the high risk of IFI in certain hematology patients, such as those with prolonged neutropenia or after allogeneic HSCT, and a wide spectrum of fungal pathogens, pre-emptive therapy L-AmB, which is fungicidal against both yeasts and molds, can be useful in ICU-admitted patients with clinical suspicion of IFIs based on one of the following: radiological findings, or cultures from non-sterile, mainly respiratory, materials, or non-culture based tests, such as GM or PCR [157]. A complete diagnostic work-up should be performed, including sampling at the site of infection, and antifungal treatment should be discontinued if the suspicion of IFI is not confirmed.

Statements

• 27. Antifungal prophylaxis, either with fluconazole to target Candida species or with posaconazole to target also molds, is recommended only in some selected high-risk populations of hematology patients (e.g., a mold-active agent in case of neutropenic patients undergoing induction chemotherapy for AML or allogeneic HSCT, or patients with graft-versus-host disease; fluconazole for patients receiving high-dose chemotherapy for aggressive lymphoma).

• 28. Patients with hematologic malignancies receiving mold-active azole prophylaxis who develop suspected or documented breakthrough IFI should receive treatment with L-AmB and promptly undergo a complete diagnostic workup.

• 29. Patients with hematologic malignancies admitted to the ICU and having IFI with no possibility for TDM of azoles and/or at high risk of azole-related drug–drug interactions should receive treatment with L-AmB.

• 30. Considering the high risk of IFI and wide spectrum of fungal pathogens in certain hematology patients (with prolonged neutropenia or after allogeneic HSCT), empirical therapy with L-AmB can be useful in patients admitted in ICU with clinical suspicion of IFIs while completing diagnostic work-up and it should be discontinued if the suspicion of IFI is not confirmed. *

Abdominal surgery

Background

Intra-abdominal candidiasis is the most common type of deep-seated candidiasis [158]. Although Candida invasion and dissemination within the abdominal cavity may occur, IAC is rarely accompanied by candidemia [90]. Thus, diagnosis of IAC without bloodstream infection may be difficult, especially in the absence of a non-culture-based gold standard method [159].

Because of the poor prognosis of IC in critically ill patients, empirical antifungal treatment is commonly administrated. However, less than 10% of ICU patients receiving an empirical antifungal therapy for suspected IC obtain a microbiological diagnosis [160]. To identify ICU patients who may benefit from the early introduction of antifungal therapy, some strategies based on clinical characteristics have been proposed. For instance, a recent algorithm differentiated patients based on the presence of septic shock [161]. Other prediction rules based the choice on the assessment of multifocal Candida [37, 162]. Despite these scores being suitable for patient bedside evaluation, they may overestimate the risk IC brings to the extensive use of antifungals.

The choice of drug for IAC is another critical issue. Currently, guidelines recommend echinocandins as the first-line treatment for IC. However, recent literature suggests their intra-abdominal penetration is limited [163,164,165]. The high plasma protein binding (> 95%) significantly affects their passive diffusion into the peritoneal fluid [65, 166]; indeed, only the unbound fraction passes from the vascular to the extravascular compartment. It is estimated that only 33% of the echinocandin dose reaches the intra-abdominal cavity [167]; moreover, some PK studies documented a low probability of PK/PD target attainment using standard dosing regimens, especially for less susceptible Candida species [149, 152]. For these reasons, some authors proposed to use of higher dosages of echinocandins for the treatment of IAC, but definitive data are lacking [168, 169]. In other terms, abdominal candidiasis could represent a hidden reservoir of resistance to echinocandins with a higher risk of failure despite adequate source control [134, 159].

L-AmB has good activity against Candida species, a low potential for inducing resistance, concentration-dependent fungicidal activity, a prolonged post-antifungal effect, and a potent anti-biofilm effect. Unlike echinocandins, available PK data evidenced that, unlike echinocandins, L-AmB did not show any significant difference in concentrations between healthy volunteers and critical patients; moreover, no decrease in Cmax or AUC was observed in patients undergoing CRRT [159]. The efficacy of L-AmB increases linearly with its concentration, showing strong fungicidal activity in deep-seated compartments such as the pleura, peritoneum, pericardium, aqueous humor, and vitreous [163]. A recently published therapeutic decision algorithm placed L-AmB as a first-line treatment in suspected or confirmed cases of IAC and sepsis/septic shock with candidemia or endophthalmitis as well as IAC and sepsis/septic shock with previous exposition to echinocandins and/or fluconazole or risk factors for N. glabratus infection [159]. In the case of echinocandin-resistant C. auris use of L-AmB, 5 mg/kg/day was proposed alone or in combination with echinocandins, as in vitro synergic activity was demonstrated [159, 170, 171].

Of course, along with antifungal therapy, an appropriate source control remains a key component of the treatment of critically ill surgical patients with IAC [172].

Statements

• 31. Diagnosis of IAC remains challenging. It is based on microscopy and culture of specimens obtained during surgery or by percutaneous aspiration. Blood cultures must be taken but might not be helpful for diagnosis due to lack of sensitivity. Non-culturable methods, BDG determination, or other tools might be used to exclude fungal etiology.

• 32. In patients with IAC, the choice of empirical antifungal therapy should be guided by host, microbiological, and epidemiological variables. L-AmB could be considered first-line therapy in cases of IAC with sepsis/septic shock, the risk for N. glabratus and C. parapsilosis infections, or previous therapy with echinocandins. *

• 33. Echinocandins could be used as a first-choice treatment in non-critically ill patients. However, recent pharmacokinetics/pharmacodynamics evidence suggested that exposure to the ascitic fluid may be suboptimal and may cause breakthrough resistance, especially in the case of non-albicans etiology. *

• 34. Combination therapy with L-AmB and an echinocandin should be considered a rescue therapy in the case of C. auris etiology.

• 35. In critically ill patients, empirical antifungal therapy for suspected IC (including those with potential abdominal origin) may be safely interrupted early according to a biomarker-driven strategy. *

Conclusions

Treatment of IFI in critical care is still challenging due to the growing number of patients at risk and the emergence of drug-resistant fungal species. With its broad-spectrum activity and major safety compared with previous formulations, LAMB may represent a suitable therapeutic choice for many clinical scenarios. For this reason, a multidisciplinary panel of 16 Italian experts developed 35 statements on the use of LAMB in ICU based on a scoping review of the most updated literature. Though the scientific debate on the place in therapy of LAMB is ongoing, this consensus document would first reach out to unmet clinical needs in critical care. Differently from current guidelines, this paper uncovers common clinical situations where LAMB may be a front-line therapy, consequently encouraging a more appropriate use.

This study has some limitations. First, this document is based on expert opinions, as evidence on the use of LAMB in the ICU population is limited. All the panel members work in Italian centers, narrowing the scope of the contents. Indeed, the use of LAMB may be precluded by the economic charge and the unavailability in some centers. Finally, new antifungals will be available in clinical practice in a short time, broadening the therapeutic armamentarium for difficult-to-treat fungal infections.

This expert opinion paper could represent a practical tool for physicians involved in the care of critically ill patients at risk for severe fungal infections. Enhancing clinical evidence on the use of LAMB in the ICU may encourage the design of high-quality prospective studies on LAMB to improve the management of IFI in the ICU.