Invasive fungal infections in acute and chronic liver impairment: A systematic review

Patients with acute and chronic liver impairment are susceptible to invasive fungal infections such as candidemia and invasive pulmonary aspergillosis as a result of cirrhosis‐associated immune dysfunction, humoral immunodeficiency, cell‐mediated dysfunction and systemic inflammation. Besides classical risk factors for invasive fungal infection, acute‐on‐chronic liver failure, corticosteroid use, gastrointestinal bleeding, and prophylactic use of antibiotics are all additional conditions which are related to the potential development of fungal infections. Therefore, high‐risk patients should be carefully followed by microbiological surveillance including cultures but also by imaging and fungal biomarkers for providing early diagnosis. Echinocandins are still the mainstay and first line antifungal therapy in cases of invasive candidiasis. Due to concerns of liver toxicity and in cases of renal impairment liposomal amphotericin B is a suitable alternative to voriconazole in patients with invasive pulmonary aspergillosis. Although, data of isavucoanzole and posaconazole use in those patients are also promising more specific studies in the subgroup of patients with liver impairment are needed. Especially, due to the late diagnosis and multiple organ dysfunction usually present in patients with liver impairment morbidity and mortality rates remain high. Based on the broad spectrum of diverse reports with varying content and quality and in some cases lack of evidence we performed a systematic review on this topic.


| INTRODUC TI ON
Invasive fungal infections (IFI) are more and more recognised as an emerging problem associated with increased morbidity and mortality rates, beyond the typical at-risk patients. [1][2][3] One of these groups are patients with liver impairment. This includes not only patients with acute liver failure (ALF) but also patients with chronic endstage liver disease (ESLD) and the sub group of patients with severe alcoholic hepatitis (SAH), which are associated with a variety of host immune dysfunctions. 4,5 Although, best investigated in ESLD, the so-called cirrhosisassociated immune dysfunction (CAID) can serve as a model for liver impairment also in ALF. In these patients, not only the innate but also the adaptive immune functions with an increased susceptibility to infections is disturbed. 6,7 CAID involves a state of immunodeficiency, and in parallel a state of persistent activation of the immune system cells, with increased production of pro-inflammatory cytokines and systemic inflammatory response syndrome (SIRS).
CAID is a multifactorial process, resulting from bacterial overgrowth, dysbiosis and increased translocation, which is responsible for a continuous (over) stimulation of immune system cells.
As recently reviewed by Albillos et al. this interaction of gut bacteria with the immune system lead to a so-called 'immune paralysis' followed by an immune dysfunction at multiple cell levels (neutrophils, monocytes, T and B lymphocytes and natural killers). 6 The incidence of infections in patients with liver impairment has been evaluated in previous therapeutic trials as part of the secondary outcomes or adverse events of the studied intervention. However, these studies are typically not designed to address the issue of infection complications, in detail IFI, in liver impairment patients.
Following these implications, IFI are underrepresented in studies but much more in clinical practice. This is why IFI are associated with an increased risk of morbidity and elevated mortality rates that may rise up between 73% and 100% in patients infected by Candida spp.
or Aspergillus spp. 4,8,9,10 Referring to the morbidity, the burden of IFI in patients with acute and chronic liver failure has been recently addressed in more focused cohort studies showing prevalence rates ranging between 2.5% and 10%. [11][12][13] In this scenario, aim of this review is to address epidemiological, clinical aspects and challenges in the early diagnosis and treatment of IFI in patients with acute and chronic liver impairment.

| Literature search
In this systematic review, we searched PubMed from January 1960 up to December 2020, for articles published in the context of fungal infections and liver disease. Focus was for both, Aspergillus spp. and Candida spp. infections: ALF; alcoholic steato hepatitis and ESLD.
The search strategy was: ('ALF and aspergillosis/candidiasis' OR 'ALF and invasive pulmonary aspergillosis' OR 'ALF and mould infection' OR 'ALF and invasive fungal disease') AND ('alcoholic hepatitis and aspergillosis/candidiasis OR SAH and fungal infection' OR 'alcoholic steato hepatitis and aspergillosis' OR 'alcoholic steato hepatitis and fungal infection) AND ('liver cirrhosis and aspergillosis' OR 'liver cirrhosis and pulmonary aspergillosis' OR 'liver cirrhosis and invasive fungal disease' OR 'acute-on-chronic liver failure and Aspergillosis' OR 'acute-on-chronic liver failure' OR 'pulmonary aspergillosis' OR 'acute on chronic liver failure and invasive fungal disease'). The same approach was used for Candida spp. and rare fungal infections.
The search was limited to English. We additionally searched reference list of included studies and relevant publications.

| Study selection criteria
For a study to be eligible for inclusion in this review, patients were required to have been diagnosed with underlying liver disease covered in this review based on underlying risk factors and clinical, laboratory, tissue histopathology and radiological findings.
Patients after liver transplantation were excluded from this review. On the one hand, these Patients are still represented in the current guidelines (solid organ transplantation), and on the other hand, the focus of this review is on patients with liver disease and not obvious immunosuppression and the risk for fungal infections.
Following our search criteria, a total of 665 articles could be identified on PubMed. Using our selection criteria and exclude doubled papers in each search category a total of 57 papers left which were included in this review.
The authors confirm that the ethical policies of the journal, as noted on the journal's author guidelines page, have been adhered to. No ethical approval was required as the research in this article related to micro-organisms.

| INVA S IVE C AND ID IA S IS (I C) IN ACUTE AND CHRONI C LIVER FAILURE
Patients with acute-or-chronic liver impairment are risk populations for developing fungemia and spontaneous Candida spp. peritonitis.
Fungemia due to Candida spp. is the most common presentation of IFI in this population, accounting for >70% of all cases of invasive candidiasis (IC). 9

| Invasive candidiasis in ALF
Patients with ALF may present a high susceptibility to Candida spp.
infections especially based on the impaired phagocytic function, the need for invasive procedures and exposition to antibiotic treatment. 6 Data are scarce revealing on the epidemiology of invasive candidiasis in ALF patients excluding the scenario of liver transplantation and acute-on-chronic liver failure (ACLF). A prospective study published in the 90's described 32% of Candida spp. infections in 50 patients with ALF. 14 Of note, this study has some limitations concerning diagnostic criteria of invasive candidiasis once cultures obtained from respiratory and mucosal samples were also considered as evidence of infection in some cases. Otherwise, if only cases with microbiologic documented invasive Candida infections were considered, the prevalence remains 8% (4/50 patients). 14 No further data has been published since then with this constellation.

| Invasive candidiasis in SAH
Severe alcoholic hepatitis is an acute hepatic manifestation typically following a heavy alcohol ingestion. SAH presents a clinical setting of liver inflammation, hepatocyte injury, and fibrosis with a varying spectrum, from mild abnormalities of liver parameters to life-threatening (acute) liver failure. 15 Beyond the typical liver injury often combined with hepatic encephalopathy, patients with SAH frequently develop renal impairment and multi-organ failure.
The treatment for SAH is until now a matter of debate, however, corticosteroids (prednisolone 40 mg/day for up to 28 days), are still the most widely used treatment option. 5,16,17,18 In addition to the high mortality rates caused by SAH itself, infections are one of the main complications, as well as one of the major causes of mortality in this setting. Patients with SAH are prone to infections, especially of bacterial origin, which are present in 30%-80% of SAH cases. 17 Only small series are published addressing IC specifically in SAH patients. The few available data assume a rate of IC in SAH varying between 2% and 9% which is associated with high mortality rates. 5,16,17,19 Considering that five out eight cases of IC were documented in ICUs, it is difficult to determine whether episodes of candidemia were strongly related to the scenario of SAH or were mainly secondary to all other risk conditions usually seen in critically ill patients demanding life support for long periods. The main epidemiological studies in this scenario were summarised in Table 1.

| Invasive candidiasis in ESLD
Most series of IC in ESLD point to ALF secondary to viral infections and alcoholic hepatitis treated with high dose corticosteroids as the main underlying conditions associated to this fungal infection. 9,11,13 The condition that most substantially increases the risk of IC is ACLF that is a syndrome characterised by acute decompensation of patients with chronic cirrhosis, multiple organ failure(s) and high short-term mortality. Fernández et al evaluated a cohort of 642 patients with ACLF or only acute decompensation without other organs failures and found that invasive fungal isolation, invasive candidiasis and invasive aspergillosis were much more likely to be documented in the first group of patients (3.9% vs 0.4%, p = .005). 10 It is consensus that most cases of IC in ESLD can be characterised as health care associated infections (65%-100%). 9,12,13 Moreover, at the time of diagnosis, most patients are usually hospitalised in ICU and exposed to several classical risk conditions for IC as invasive medical procedures, antibiotics, steroids, parenteral nutrition, and haemodialysis.  In contrast, spontaneous peritonitis is a severe complication of advanced liver cirrhosis that can be found in up to 12% of patients with ESLD. 12 Fungi are implicated as the aetiologic agent of spontaneous peritonitis in 3%-10% of cases, exhibiting higher mortality rates when compared to bacterial peritonitis (35%-100% vs 10%-50%). 4,11,13,21,22,23,24 The most comprehensive review on spontaneous fungal peritonitis (SFP) was published by Tariq in 2019 and reported a total of 82 cases described in the literature so far. Candida species, predominantly Candida albicans was the main fungal pathogen described (48%-81.8%) followed by Candida glabrata (6.66%-20%).
Cryptococcus neoformans was the second most common genera of yeasts causing SFP. 24 The exposition of patients to broad spectrum antibiotics is a major inducer of gastrointestinal dysbiosis and increases substantially their risk of developing IC.
Antibiotic prophylaxis is frequently prescribed for ESLD patients with gastrointestinal bleeding, with previous SBP and patients with ascites exhibiting fluid protein <1.5 g/dl associated to either impaired renal functional or liver failure. In this context, the reduction in the intestinal bacterial microbiota results in significant fungal colonisation that may lead to translocation across the damaged gastrointestinal tract mucosa into the peritoneal cavity. The immunosuppression and malnutrition that are characteristically found in cirrhotic patients are both conditions that may promote fungal translocation and SFP. [25][26][27][28] In Table 2 epidemiological and clinical aspects related to candidemia and episodes of SFP in patient with ESLD as reported by studies describing at least 10 cases are described.

| INVA S IVE PULMONARY A S PERG ILLOS IS IN ACUTE AND CHRONI C LIVER FAILURE
Approximately 43%-80% of invasive pulmonary aspergillosis (IPA) cases appear in nonhaematological malignancy patients. 3 Typically, these patients do not fulfil the EORTC/MSG criteria for invasive aspergillosis, radiological findings are unspecific, and the performance of fungal biomarkers may be suboptimal when only serum samples are tested.
Lack of awareness, delayed diagnosis and treatment resulting in dramatically high mortality rates of up to 90%-100%. 5,16,17 Even more, prophylactic strategies are until now neither established nor systematically investigated. In the following, the role of IPA in acute and chronic liver failure will be reviewed.

| Invasive pulmonary aspergillosis in ALF
The incidence and clinical relevance of IFI, especially of IPA in ALF is unknown and might be underdiagnosed. There are only small reports focusing on IPA in ALF, as described in Table 3. Even in the largest cohort study of Zhang et al. 29 from 2017 most patients with IPA are not suffering from ALF but from ACLF which will be discussed in the following.
Based on these data a general recommendation for screening of these patients could not be given, it should be recognised as a potential complication especially in cases of pulmonary worsening under antibiotic treatment.

| Invasive pulmonary aspergillosis in SAH
Although, associated with tremendously high fatality rates, little is known about the incidence of IPA in SAH. Based on selected reports, incidences vary between 16% and 42% (see Table 1).
Controlled postmortem or autopsy studies were not reported in patients with SAH and IPA.
Two important case series addressed IPA in SAH. A prospective study that followed 94 biopsy-proven severe AH episodes along 3 months reported fifteen cases of invasive aspergillosis (six proven, eight probable, and one possible) in this setting. The severity of this condition resulted in 100% of mortality in patients not eligible for liver transplantation, despite antifungal treatment. 16 Another retrospective study of biopsy-proven SAH in ICU, found an even higher incidence of 40% of IA, with 100% mortality. 5 The high-reported mortality rates, above all in the ICU cohort, may also be explained by relevant comorbidities such as respiratory failure, acute decompensation of the liver or other coinfections beyond the SAH.
Beyond CAID caused immunodeficiency, superimposed SAH seems to worsen the immune status by a more markedly immunosuppressive profile of T lymphocytes (higher interleukin-10 expression and lower interferonγ production) due to overexpression of inhibitory receptors (PD1, PDL1, TIM3 and galectin-9) and reduced neutrophil antimicrobial activities. Thus, it is reasonable to expect a higher incidence of infection in SAH than in ESLD. 15 On the other hand, a reasonable consideration is that cortico-  Table 4).
Reported incidences vary between 0.3% and 14%, whereas the most studies report IPA rates of 5%-6% of observed patients. 33,34 Studies with higher incidences typically report about critically ill patients. 35 Risk factors for IPA in this population are a higher CHILD Pugh score (most patients CHILD C), elevated MELD and a higher ACLF grade. IPA in ESLD is typically associated with critical illness including mechanical ventilation with a higher proportion of patients with need of renal replacement therapy. COPD is in nearly 30% of these patients a typical comorbidity. This is in the line with the findings of Prattes et al. 34 In this study, only 1.3% of the included patients had probable IPA. However, none of these patients were critically ill or fulfilled the criteria mentioned above.
High mortality rates of up to 100% in ESLD patients with IPA may be explained by CAID but furthermore, by the delayed diagnosis due to a low index of suspicion and often the misinterpretation of fungal samples as colonisation.

| OTHER FUNG AL INFEC TI ON S IN PATIENTS WITH LIVER IMPAIRMENT
Besides IC and Aspergillosis, considering the impaired phagocytic function usually associated to patients with advanced liver disease, cases of cryptococcosis and endemic mycosis have also been reported in this population.
Infections caused by Cryptococcus spp. have been described mostly in patients with ESLD.
The recent systematic review regarding SFP described

Cryptococcus neofarmans as the second main fungal pathogen, after
Candida spp. 24 One of the main manifestations of cryptococcal disease in this scenario are spontaneous peritonitis, representing at least 30% of cases, followed by meningitis and pulmonary disease. 36,37 Fifty per cent of patients with spontaneous peritonitis had simultaneously positive blood cultures for Cryptococcus spp. 36,38 Moreover, decompensated liver cirrhosis is described as a risk factor for developing cryptococcal invasive disease, with an adjusted odds ratio of 23.8 and 5.3 in the Taiwanese and United States casuistics, respectively. 37,39 The authors mentioned that cryptococcosis in patients with cirrhosis was associated with high mortality rates (80%) and rapid progression to death, with median survival of 6 days in an American retrospective cohort. 36 Based on this sporadic reports and the small evidence prophylactic regimes, especially for PCP may be used as an individual approach.

| D IAG NOS TI C APPROACH FOR IFI
The low clinical response to antifungal therapy documented in most case series is probably related to a combination of factors including the severity of underlying conditions, low suspicion of fungal infections by clinicians, limited sensitivity of available diagnostic methods and late diagnosis and initiation of antifungal treatment (see Tables 1-5). Indeed, checking 16 papers 4,5,11,12,14,16,19,20,31,33,36,38,40,50,51,52 where details about the chronology of the diagnosis were provided This is in the line with some evidence suggesting that this population is more likely to generate false-positive results, for example of 1,3 Beta D Glucan (BDG) assays, a pan-fungal antigen test that has been widely used to early detect patients with invasive candidiasis and is also used in cases of IPA. 55,56 Otherwise, these antigen tests may be useful to exclude the possible fungal infections considering its high negative predictive value, including patients with suggested IPA. A single-centre study conducted by Verma et al., 33  Nevertheless, no cirrhotic patient was included in this casuistic. 56 Beyond biomarkers, major diagnostic criteria for spontaneous peritonitis is a polymorphonuclear leukocyte count >250 cells/mm 3 in ascitic fluid. It is important to mention that differences in cytology and biochemical analysis of peritoneal fluid fail to predict fungal infections TA B L E 5 Summary of cases series reporting 10 or more episodes of cryptococcosis and endemic mycosis in patients with chronic liver failure from any other pathogens, and coinfections with bacteria can be present in up to 75% of cases. 11,21 Aiming to improve fungal infection diagnosis, inoculation of ascitic fluid into aerobic and anaerobic blood culture bottles and standard microbiological testing should be performed. 57 Even more challenging, as reported in Verma's study and other series, is to achieve laboratorial criteria required for the categorisa-  To date, the antifungal agents licenced for the first line treatment of IPA include voriconazole, isavuconazole and liposomal amphotericin B. However, as also stated in the latest ESCMID-ECMM-ERS guidelines, liposomal amphotericin B is usually the first therapeutic option in patients with liver insufficiency. 60 This recommendation is based on the experiences of drug-drug interactions and toxicities in patients with liver impairment and voriconazole treatment.
In contrast, in two small studies, the treatment with liposomal amphotericin B in patients with liver impairment presented good tolerability. 58,86 Overall, Isavuconazole opens a new possibility for treatment of IA, with fewer toxicities than polienics and voriconazole, presenting predictable pharmacokinetics. 87 Moreover, as recently published by Maertens et al. 88

| CON CLUS ION
Patients with acute-or-chronic liver impairment are prone to infections, mainly of bacterial origin, however, IFI are emerging, are underdiagnosed and are associated with high morbidity and mortality rates.
Based on pathophysiological implications rooted in complex interactions between liver impairment and the immune system, a high suspicion to IFI is warranted. A prompt and aggressive diagnostic strategy, including traditional cultures of specimens, biomarkers and imaging is needed to reveal these opportunistic infections. This strategy includes also an adequate antifungal treatment which was associated with a decreased risk of 30-day mortality. The role of antifungal prophylaxis is currently unknown and should be investigated in clinical trials.