Invasive candidiasis in critically ill patients: Fundamental concepts and future directions

Introduction

Candida spp. are a group of eukaryotic yeasts, several of which are known to cause significant human infections. Among the most clinically relevant are C. albicans, C. parapsilosis, Pichia kudriavzevii (formerly C. krusei), Nakaseomyces glabratus (formerly C. glabrata), Candida tropicalis, and Candidozyma auris (formerly Candida auris); some species were renamed after phylogenetic studies showed that they belong to other genera.^1,2

Candida infections include superficial skin and mucocutaneous infections (which do not usually cause severe complications) and invasive infections.³ Invasive infections caused by Candida species are known as invasive candida infections (IC).² IC is further classified into two categories: deep-seated infections (e.g., intra-abdominal candidiasis [IAC]), in which Candida invasion is demonstrated by culture or histology in typically sterile sites, and bloodstream infections, commonly known as candidemia.²

Candidemia represents about 20% of bloodstream infections in critically ill patients⁴ and is associated with approximately 40% mortality.^5,6 Moreover, despite advances in treatment and updated clinical guidelines, patients with candidemia often experience prolonged intensive care unit (ICU) stays⁷ and have an increased risk of complications related to infection seeding in distant organs.²

This article will review fundamental concepts necessary to understand the development of the principal forms of IC in ICU patients —candidemia and IAC— along with practical and updated approaches to their diagnosis and treatment. Given the high prevalence and associated mortality of Candida infections in the ICU, a basic understanding of the pathogenesis, diagnosis, and treatment of these infections is important for the ICU clinician.

Literature search

A literature search was conducted in PubMed/MEDLINE, Scopus, and Embase using the MeSH terms “Candida” and “Candidiasis”, including the subheadings “Epidemiology,” “Physiopathology,” “Microbiology,” and “Diagnosis.” Additional keywords included “Critical care” and “Antifungal agents.” These terms were combined using Boolean operators, and the search was focused on studies involving critically ill patients with particular emphasis on the past 10 years; the search was current through October 2025.

Epidemiology

IC is a frequent complication in ICU patients, with an estimated incidence of 7 episodes per 1000 ICU admissions.⁸ However, its true incidence may be underestimated due to difficulties with diagnosis, discussed below.⁹ Candidemia is the leading manifestation of IC,⁸ and its incidence varies across countries. Focusing on the United States (U.S.), the incidence has been reported at 9.8 per 1,000 ICU admissions¹⁰ and was stable over the decade preceding the COVID-19 pandemic.¹¹ However, during the pandemic, an increase in cases of candidemia was reported.^12,13

After candidemia, IAC is the most common site for deep-seated infections, with an incidence of 40 cases per 100,000 hospitalizations. IAC is a heterogeneous condition but most commonly manifests as intra-abdominal abscesses (55%), followed by secondary peritonitis (33%) and primary peritonitis (5%).¹⁴ About 14% of IAC patients have concomitant candidemia, and therefore, patients with IAC should be assessed for bloodstream dissemination.¹⁵ The 30-day IAC mortality rate in ICU patients is 38%, and source control in drainable collections improves survival.^14,15

Regarding species distribution in candidemia, C. albicans is the most frequently isolated species, representing over 50% of cases, followed by C. parapsilosis, N. glabrata, and C. tropicalis.³ A similar microbiological pattern is observed in IAC, where C. albicans remains the main species.¹⁵ However, in the past two decades, studies have reported a shift, with an increase in cases of non-albicans Candida species, particularly N. glabrata and C. parapsilosis,^3,16 possibly linked to the broad use of antifungals like fluconazole for prophylaxis.³ Because some non-albicans species have higher resistance rates to azoles than C. albicans, using azoles can create selective pressure that favors their growth.¹⁷ In parallel, acquired antifungal resistance is increasing among non-albicans species; of particular concern, clonal outbreaks of fluconazole-resistant C. parapsilosis have been reported globally, with horizontal transmission facilitated by its ability to persist in the environment despite decontamination efforts, and can occur among patients without previous azole exposure.^18-20 Although less common, echinocandin resistance has been reported in N. glabrata and P. kudriavzevii, potentially limiting therapeutic options.^21-23

Among Candida species, C. auris is one of the fungi listed by the World Health Organization (WHO) as a “critical” priority pathogen due to the increasing number of cases globally, its capacity to cause outbreaks, and its potential for multidrug resistance.²⁴ Importantly, 90% of C. auris isolates are resistant to fluconazole, 30% are resistant to polyenes (e.g., amphotericin-B), and 5% are resistant to echinocandins.²⁵ In the U.S., over 4,500 cases have been reported, and cases continue to rise.²⁶ Similar to fluconazole-resistant C. parapsilosis, C. auris establishes environmental reservoirs on hospital surfaces and within colonized patients, leading to difficult-to-control, endemic infections in healthcare facilities.²⁷ The changing epidemiology and emergence of drug resistance in non-albicans species highlight the importance of both local and nationwide surveillance of Candida infections.

Pathogenesis

Candida spp. are common commensal organisms found on the skin ²⁸ and gastrointestinal tract (GIT) of healthy individuals.²⁹ However, not all Candida species are part of the normal microbiome. In the GIT, C. albicans and C. parapsilosis are common, whereas C. tropicalis and N. glabrata are less frequently found.²⁹ Furthermore, C. auris has not been described as part of the healthy gut microbiome.²⁹

In critically ill patients, factors such as antibiotic exposure, indwelling devices, and immunosuppression promote further Candida colonization and overgrowth.³⁰ After exposure to those predisposing conditions, Candida can progress from a commensal organism to an invasive pathogen. This process can be divided into three stages: initial colonization, subsequent tissue invasion, and dissemination (Figure 1).^31,32

Figure 1:

Steps from Candida colonization to invasion. In the gut, a healthy microbiome is characterized by diverse microbiota and metabolites such as short-chain fatty acids (SCFAs). When an environmental change occurs, such as during antibiotic use, the microbiome disruption allows Candida to overgrow. Then, Candida adheres to the epithelium and transforms into hyphae. Later, the hyphae penetrate the epithelial barrier, leading to invasion and subsequent vascular dissemination. During this phase, the hyphae transform back into yeast. This process triggers the immune response, activating TH17 lymphocytes and neutrophils.

In the skin figure, the first image represents that C. auris is not part of the normal microbiota and is acquired after environmental contact. In the second image, C. auris colonizes the skin, forming biofilms and altering the microbiome. In C. auris colonized patients, an association with gram-negative colonization has been described, which is also represented in the figure. After evading macrophage and neutrophil responses and entering through a skin barrier breach, C. auris can disseminate.

In in vitro models, Candida colonization begins with the adhesion of its yeast form to the epithelium, a process mediated by surface adhesins.³³ A key feature of C. auris colonization is its ability to persist on the skin and in the nares for long periods.²⁷ This is due to its ability to form a biofilm, a structure composed of microorganisms and extracellular matrix, in deeper skin layers and hair follicles. This structure protects the fungi from the host’s immune system, leading to a higher fungal load.³⁴ In the GIT, C. albicans may adopt alternative morphological states, including the gastrointestinal-induced transition cell type (GUT), an adapted metabolically active cell that uses nutrients in the distal GIT, which favors Candida survival.^35,36

After Candida colonizes a surface in the oral epithelium, the yeast transforms into its hyphal form and invades the epithelium.³³ In the intestinal mucosa, the normal conditions of the gut microbiome suppress hyphal morphogenesis. However, in conditions where the microbiome's homeostasis is lost, Candida can form hyphae that invade the epithelium, ultimately leading to bloodstream translocation. This process can be accelerated by a loss of the mucosal barrier during chemotherapy or other conditions causing epithelial damage.³⁷ After passing into the bloodstream, Candida spp. primarily exists as yeast, the cell type associated with dissemination. However, the transformation from hyphae back to yeast following translocation remains poorly understood.³¹ At the same time, during invasive candidiasis, the immune response promotes T cell differentiation; T cells release IL-17A and interferon-gamma (IFN-γ), cytokines that activate neutrophils and induce the release of antimicrobial peptides in the epithelium.³³

Risk factors for candidemia

ICU patients frequently have multiple risk factors for candidemia (Figure 2). Many of these risks are associated with common ICU interventions such as parenteral nutrition, renal replacement therapy, mechanical ventilation, and blood transfusions.³⁸ Also, the presence of indwelling devices such as intravascular lines, urinary catheters, and ventricular assist devices increases the risk, as Candida species can form biofilms on these surfaces, and they provide direct access past normal skin and mucosal barriers.²

Figure 2.

Risk factors for candidemia. (1) Exposures during intensive care unit (ICU) ICU stay, including invasive procedures, mechanical ventilation, intravascular catheters, and prolonged hospitalization. (2) Environmental factors, like surface contamination and healthcare worker-mediated transmission. (3) Microbiome-related factors, such as microbial disruptions caused by broad-spectrum antibiotics and antifungal treatments. (4) Host-related factors, such as immunosuppression (e.g., neutrophil dysfunction) and advanced age

Candida colonization is the first step toward invasion and an independent risk factor for candidemia.³⁸ Because of this, colonization has been included in clinical tools used to predict infection, like the Candida Colonization Index ³⁹ and the Candida Score.⁴⁰ However, because of its low positive predictive value (PPV),^39,40 colonization should be considered alongside other clinical factors to improve the accuracy of candidemia prediction.

Antibiotics are commonly administered in the ICU, representing an additional risk factor for Candida colonization and infection.³⁸ Antibiotics deplete commensal gut bacteria that suppress fungal growth via antagonistic interactions.³¹ Commensal bacteria also produce short-chain fatty acids, fermentation metabolites that inhibit Candida overgrowth and hyphal formation due to changes in the gut's pH. There are likely other unidentified mechanisms that link commensal GI microbes to suppression of Candida overgrowth.⁴¹

Sepsis is another important risk factor for subsequent candidemia.^3,38 Sepsis leads to GIT dysbiosis and increased intestinal permeability, facilitating microbial translocation into the bloodstream.⁴² In addition, in the late stages of sepsis, an imbalance leading to immune suppression occurs, which may increase the risk of candidemia.⁴³

Other than sepsis, immune-compromising conditions such as diabetes, liver disease, solid neoplasms, and hematological malignancies are also known to increase the likelihood of candidemia.³⁸ Moreover, newer immunosuppressive therapies have made additional patient groups more vulnerable, such as those receiving monoclonal antibodies that target IL-17A.⁴⁴ Among specific immunosuppressed populations, solid organ and stem cell transplant recipients are at risk even with standard antifungal prophylaxis. These infections often involve non-albicans Candida and may occur outside typical risk periods.^45,46

Diagnosis

Clinical signs of invasive candidiasis include those related to the infection itself or to the organ compromised in deep-seated infection.⁴⁷ Signs of candidemia are similar to those of bacterial sepsis, making early recognition challenging. Early recognition therefore requires a high index of suspicion in patients with signs of sepsis and risk factors for Candida infection, especially in those with underlying immunosuppression.⁴⁸ IAC manifestations include signs of peritonitis or intra-abdominal collections, and IAC should be suspected in patients with conditions such as gastrointestinal perforation, anastomotic leaks, and recent antibiotic use.⁴⁹

Diagnostic methods for IC include cultures, biomarkers, and molecular techniques, summarized in Table 1. For candidemia, blood culture is the most conventional diagnostic test and the gold standard. However, its sensitivity has been reported to be less than 50%.⁵⁰ This estimate comes primarily from autopsy studies, which were limited by small sample sizes and did not confirm every case with histology.⁵¹ Experimental studies have shown that blood culture sensitivity is likely higher than 50%, and true sensitivity is not completely clear.⁵²

Table 1.

Performance of classic invasive candidiasis clinical prediction scores and diagnostic tests. List of abbreviations: AUC (area under the curve), CFU (colony-forming unit), CVC (central venous catheter), FDA (Food and Drug Administration), NPV (negative predictive value), PCR (polymerase chain reaction), and PPV (positive predictive value).

Diagnostic test	Performance	Comments
Blood culture	Sensitivity 50%50	Limit of detection 1 CFU/ml Time to positivity could be >48 hours Lower sensitivity in deep-seated candidiasis
Tissue culture in IAC	Sensitivity 50% for hepatic candidiasis50	May be affected by antifungal therapy Requires an invasive procedure to be obtained
β-D-glucan for candidemia	Sensitivity 81% Specificity 60% 56	False positives: Dialysis with cellulose membranes, albumin or blood products, immunoglobulin therapy, conditions that increase gut permeability (e.g., septic shock, abdominal surgery) False negatives: Infections caused by non-albicans species such as C. auris or C. parapsilosis, prior antifungal therapy, infections in sanctuary sites (i.e., central nervous system and eye), or poorly vascularized sites (e.g., bone)103
β-D glucan for IAC	Serum β-D-Glucan58 Sensitivity 59% Specificity 63% PPV 55% – NPV 66% Peritoneal B-D-Glucan58 Sensitivity 89% Specificity 76% PPV 53 - NPV 82	False negatives: Infections with low Candida inoculum Exposure to antifungals
PCR for IAC	Sensitivity 64%55 Specificity 89% PPV 61% - NPV 90%	In-house tests Better performance in high-risk patients (e.g., recurrent abdominal surgery, anastomotic leaks, severe pancreatitis, multifocal Candida colonization)
Multiplex PCR (Biofire)	Sensitivity 92%64 Specificity 95%	Requires a positive blood culture first Identifies C. auris, but not C. dubliniensis
T2 Candida panel	Sensitivity 91%62 Specificity 99%	FDA approved in 2014 It can detect as few as 1 CFU/mL Results in 5 hours The result is positive or negative for three groups of species: C. albicans/C. tropicalis, N. glabrata/C. krusei, and C. parapsilosis; meaning at least one species in each group was detected
Predictive scores	Performance	Comments
Ostrosky-Zeichner	PPV 5%67 NPV 98% AUC 0.70	Criteria: Total parenteral nutrition, hemodialysis, major surgery, pancreatitis, immunosuppressants, and broad-spectrum antibiotics
Candida score	PPV 16%104 NPV 98% AUC 0.80	Criteria: Parenteral nutrition, severe sepsis, multifocal Candida colonization

In addition, factors like blood culture volume and the use of fungal blood culture media may affect culture performance. Obtaining 2 or 3 sets of blood cultures, each with 20 mL of blood, can increase the likelihood of isolating Candida.⁴⁷ Also, using fungal-specific blood culture media has been associated with faster time to positivity compared to bacterial media.⁵³ Considering that time to positivity in candidemia can be as long as 2-3 days in some cases, decreasing time to diagnosis may lead to earlier therapy and improved patient outcomes.⁵⁴

In IAC or deep-seated candidiasis, Candida inoculum in the blood may be too low for detection by blood cultures. Therefore, samples from sterile sites like biopsies or fluid aspirates are needed, although these too may have low yield.^47,50 To overcome the limitations of conventional cultures in IAC, polymerase chain reaction (PCR) testing on a sterile site sample has been proposed as a faster and more sensitive method.⁵⁵

Furthermore, diagnostic strategies have expanded to include biomarkers to stratify the likelihood of candidemia in ICU patients. The most extensively used is 1,3-β-D-glucan, which may have a sensitivity of 81% and a specificity of 60% for ICU patients at risk of invasive candidiasis and candidemia.⁵⁶ However, many conditions can cause false positives and negatives with this test, which are listed in Table 1.⁵⁷ Therefore, two consecutive values have been recommended, and the results should be interpreted cautiously.⁴⁷ 1,3-β-D-glucan has also been investigated for its potential in ruling out Candida peritonitis because of its negative predictive value (NPV). However, there is not enough evidence to support its routine use in IAC.⁵⁸

Molecular tests that detect Candida DNA from blood samples have been developed to support early candidemia diagnosis. Because most of these tests are developed by local laboratories, their broader implementation is restricted and current guidelines do not strongly recommend them.^47,59,60 Among the molecular assays available, the T2 Candida panel is recognized as a fully automated and clinically validated test that can be performed directly on whole blood and provides results in less than five hours.⁶¹ The T2 Candida panel has high sensitivity (91%) and specificity (99%).⁶² These characteristics support its use in diagnosing candidemia and suggest it could assist in decisions to stop empiric antifungal therapy.⁶² As an additional advantage, prior antifungal use does not seem to affect its accuracy.⁶³ However, the cost of this method limits its use, and it has not shown an impact on reducing hospital stay or in-hospital mortality compared to standard blood cultures. Its cost-effectiveness may depend on the local prevalence of candidemia and its proper use along with a stewardship program to reduce antifungal use.⁶²

In addition to molecular tests that detect Candida DNA directly from blood samples, another group of assays is available to identify the species more rapidly from blood cultures that have already tested positive. However, these still require time for blood cultures to become positive. This includes the BioFire Film Array system by BioMerieux, which amplifies and identifies DNA from bacteria and yeasts. One of its benefits is the ability to identify the main Candida species, including C. auris, although it does not detect C. dubliniensis.⁶⁴ Another test in this group is SepsiTyper, which extracts proteins directly from positive blood cultures to identify species in about six hours.⁶⁵

Candidemia prediction: classic scores and new models

Given the mentioned caveats of candidemia diagnostic tests, diagnosis and treatment can be delayed. Several scoring systems have been developed to assist clinicians in identifying at-risk patients who may benefit from early antifungal therapy. These prediction scores share factors such as central venous catheters, dialysis, parenteral nutrition, recent antibiotic exposure, Candida colonization, use of immunosuppressive therapies, and abdominal surgeries.⁶⁶ Between them, the Ostrosky-Zeichner⁶⁷ and Candida scores⁴⁰ are widely known clinical tools (Table 1). However, their performance indicates they are more effective at ruling out than ruling in candidemia and therefore may not help clinicians decide which patients should receive empiric antifungals.

Machine learning (ML) is emerging as a potential tool to improve candidemia prediction (Figure 3).⁶⁸ ML models can incorporate more variables than traditional scores, and AUCs can reach nearly 0.9. Yet, sensitivity and specificity are not consistently high.^69,70 Moreover, the studies using ML to diagnose candidemia are limited by small patient cohorts and their reliance on biomarkers such as 1,3-β-D-glucan or procalcitonin, which may not be available when the score is used at the bedside.^68-70

Figure 3.

Machine learning for candidiasis prediction. Data from ICU patients, including clinical signs, laboratory results, and relevant medical history details, are used to train the machine learning models. First, the data is divided into two sets: One to build the model and the second to test how well it works. Models like decision trees, linear models, and neural networks analyze the data, select the most important factors, and use them to construct the prediction score for candidiasis.

Further evaluation during candidemia

Once candidemia is diagnosed, a careful physical exam is essential to identify possible sources and sites of dissemination.⁴⁷ In ICU non-neutropenic patients, if a central venous catheter (CVC) is in place, it should be removed promptly since it may be the source of infection or become colonized during infection, allowing Candida to form biofilms.^59,60,71 Failure to remove CVCs has been linked to higher mortality.⁷¹ However, in neutropenic patients, the gastrointestinal tract may more commonly be the source of candidemia due to mucositis, and the decision to remove the catheter is made on a case-by-case basis.⁵⁹

Chorioretinitis and/or endophthalmitis are complications in less than 10% of candidemia patients.⁷² The IDSA (Infectious Diseases Society of America) and ESICM/ESCMID (European Society of Intensive Care Medicine and European Society of Clinical Microbiology and Infectious Diseases) guidelines recommended a retinal examination for all patients with candidemia.^59,60 However, the ECMM/ISHAM/ASM guidelines (European Confederation for Medical Mycology, International Society of Human and Animal Mycology, and the American Society for Microbiology) recommend fundoscopy for only specific patient groups, including those with ocular symptoms, immunocompromised status, persistent candidemia, or an inability to communicate symptoms.⁴⁷

Likewise, the prevalence of endocarditis is about 6% in hospitalized patients with non-persistent candidemia, and patients with mechanical valves are at the highest risk.⁷³ Contrary to ESICM/ESCMID guidelines that recommend routine echocardiography in candidemia patients,⁶⁰ the ECMM/ISHAM/ASM guidelines have recommended its use for patients with signs of endocarditis, persistent candidemia, valvular disease, or a cardiac implantable device.⁴⁷

Routine ultrasound or computed tomography (CT) is not recommended for non-neutropenic ICU patients with candidemia.⁶⁰ However, in patients with prolonged neutropenia and candidemia, abdominal imaging should be considered due to the risk of chronic disseminated candidiasis.⁷⁴ Similarly, when evaluating patients with risk factors for IAC, a CT should be considered since abscesses can be present in 15% of cases.⁴⁹

Treatment

1. Antifungal treatment

Prevention and treatment with antifungals can be divided into four phases: prophylaxis, preemptive, empiric, and targeted therapy. Prophylaxis is an antifungal prescribed to prevent an invasive fungal infection.⁷⁵ This strategy is recommended for severely immunocompromised patients such as those receiving induction chemotherapy, allogeneic hematologic stem cell transplant, and solid organ transplant.⁴⁷ In patients with recent surgery, another at-risk group in the ICU, prophylaxis has not been shown to reduce mortality or incidence of invasive candidiasis.⁷⁶

Preemptive strategies involve starting antifungal treatment in patients without clear signs of infection, for example based on 1,3-β-D-glucan testing. This approach is not recommended, as it may lead to unnecessary antifungal use without improving outcomes.⁶⁰

Empiric treatment refers to starting antifungals in patients with signs of infection before a definitive candidiasis diagnosis is made. Empiric treatment was evaluated in the landmark EMPIRICUS trial⁷⁶ and did not reduce mortality among patients with high risk for candidemia. Current IDSA and ESICM/ESCMID guidelines recommend empiric antifungal treatment only for patients with septic shock and risk factors for candidemia, although there is no high-quality data to determine which patients should receive empiric antifungals.^59,60 Also, the IDSA guidelines propose a low threshold for empiric therapy in patients without shock who have persistent fever without another clear cause based on clinical evaluation, biomarker results, and colonization status.⁵⁹

In ICU patients receiving empiric therapy, 1,3-β-D-glucan can be incorporated into a diagnostic-driven approach.^77-79 Given its high negative predictive value, a negative result can support discontinuing antifungals in the absence of microbiological evidence of candidiasis, serving as a stewardship strategy to avoid unnecessary treatment without affecting mortality.^77-79 Importantly, this approach has not been validated in immunocompromised patients, such as patients with neutropenia or receiving immunosuppressive therapy.^77-79 In addition,1,3-β-D-glucan may help distinguish Candida from bacterial infection.⁸⁰

Once candidiasis is diagnosed, the treatment is called targeted therapy.⁶⁰ Guidelines recommend echinocandins (e.g., micafungin) as the first-line treatment for both empirical and targeted therapy.^47,59,60 This is supported by studies showing that echinocandins are associated with lower mortality and have a better safety profile than polyenes (i.e., amphotericin-B) and azoles.^81,82 Additionally, echinocandins have broader activity against antifungal-resistant Candida species than azoles.⁴⁷ Comparative studies have not shown significant differences in survival and infection recurrence among the available echinocandins (caspofungin, anidulafungin, micafungin, and rezafungin).^83-85

Once the species and antifungal susceptibility are known, treatment can be adjusted. In patients with fluconazole-susceptible isolates, de-escalation to fluconazole in the first 5 days of therapy is recommended.⁸⁶ This switch is recommended in hemodynamically stable patients with evidence of candidemia clearance, adequate source control, and tolerance of oral therapy.⁴⁷

During candidemia, guidelines recommend that blood cultures be performed every day or every other day to confirm that the blood cultures have cleared. The duration of therapy is 14 days after documented clearance and resolution of symptoms.^59,60 Other treatment recommendations for candidiasis are explained in Table 2. Recently, the concept of ‘uncomplicated candidemia’ has been proposed to identify patients at low risk of a complicated course, who might benefit from shorter therapy.⁸⁷ Proposed criteria include absence of relevant immunosuppression; adequate source control; absence of extra-abdominal, deep-seated candidiasis; clinical response with clearance of blood cultures within 5 days; and availability of antifungal susceptibility testing.⁸⁷ The evidence for short treatment courses in candidemia is scarce, although current data suggests no increased association with mortality.^88,89 Shorter courses may minimize antifungal exposure, reduce the risk of resistance emergence, and shorten hospitalization. However, further research is needed.⁸⁸

Table 2.

Summary of the guidelines for management of Invasive candidiasis. List of abbreviations: BID: twice daily, CSF: cerebrospinal fluid, PO: per os (by mouth), QD: once daily.

Invasive candidiasis	Antifungal management	Other treatment recommendations
Candidemia	Empirical treatment 47,59: Caspofungin 70 mg IV loading dose, then 50 mg IV QD Micafungin 100 mg IV QD Anidulafungin 200 mg IV loading dose, then 100 mg IV QD Rezafungin 400 mg IV once a week in week 1, then 200 mg IV once a week from week 2 Consider switching to fluconazole within 5 – 7 days if the patient is clinically stable and non-neutropenic, has negative blood cultures, the source is controlled, and isolates are susceptible to fluconazole. Once the patient tolerates oral intake, oral fluconazole may be used. Fluconazole 400 – 800 mg IV or PO daily	Order blood culture every other day until negative Treat for 14 days after blood culture clearance47,59 If CVC is in place, remove it as early as possible Identify and manage the source of candidemia as early as possible
Intra-abdominal candidiasis	Empirical treatment with echinocandins and switch to fluconazole under the same conditions as above47,59	The duration of therapy is 14 days if the source has been controlled47,59,60 Source control, including abscess drainage, leakage repair, and removal of any indwelling device, is essential
Endocarditis	Initial treatment 47,59: Caspofungin 150mg daily Micafungin 150mg daily Anidulafungin 200mg daily Or Liposomal Amphotericin B 3 – 5 mg/kg QD with or without Flucytosine 25 mg/kg four times a day	Involve cardiothoracic surgery. Duration of treatment is a minimum of 12 weeks47,59
Intra ocular candidiasis	If endophthalmitis options for first line include47,59: If susceptible isolates: Voriconazole 6 mg/kg BID on day 1, then 4 mg/kg BID from day 2, perform TDM to ensure appropriate levels Fluconazole 800mg (12mg/kg) loading dose, then 400 –800mg (6mg/kg) QD If azole resistance: Liposomal Amphotericin B 3 mg/kg QD If macular or vitreous involvement is present, consider starting Intravitreal Amphotericin B deoxycholate or voriconazole47,59	The duration of therapy is 4 -6 weeks, depending on ophthalmological examinations47,59 Consider intravitreal antifungal therapy and vitrectomy.
Central nervous system	Start treatment with liposomal amphotericin 3-5 mg/kg QD plus flucytosine 150 mg/kg QD47 If the liposomal form is unavailable, then deoxycholate amphotericin B 0.5 – 1 mg/kg QD plus flucytosine 150 mg/kg QD After source control, consider fluconazole 800 mg BID plus flucytosine 150 mg/kg QD	Source control, if feasible, and removal of any infected devices are essential to the treatment The duration of therapy must be individualized. Monitor symptoms and CSF until normalization

2. Candidozyma auris

Regarding C. auris, echinocandins remain the preferred first-line therapy before susceptibility results are known, since resistance to this class is rare and definitive therapy should be guided by susceptibility testing.⁹⁰ If candidemia persists after five days and there is no clinical response, switching to liposomal amphotericin B should be considered. Due to the high rate of azole resistance, oral treatment is not recommended for C. auris.⁹⁰

Combination therapy of antifungals in the case of C. auris has been proposed to enhance fungal clearance and improve the evolution of resistance. For example, when flucytosine was used in combination with amphotericin B and echinocandins in animal models, it showed synergy and improved survival. However, there is minimal clinical data concerning the possibility of developing flucytosine resistance.⁹¹

3. Therapeutic drug monitoring

Antifungals' pharmacokinetic (PK) parameters (including absorption, distribution, metabolism, and excretion) are affected by conditions common in the ICU, such as hypoalbuminemia, renal or hepatic failure, and fluid overload. These conditions are not considered in standard dosing adjustments for echinocandin and amphotericin; however, specific dose adjustments are advised for fluconazole and voriconazole in the presence of renal or hepatic impairment, respectively.⁹²

In ICU patients, the desired therapeutic drug levels (TDL) of echinocandins are determined by the relationship between drug exposure, quantified as the 24-hour area under the concentration-time curve (AUC), and the minimum inhibitory concentration (MIC) of each Candida species.⁹³ For infections caused by species with high MICs, it has been suggested that higher doses are needed to reach the TDL.⁹⁴ However, it is unclear whether increasing the dose for these species improves patient outcomes. As such, no major guidelines endorse species-specific echinocandin dose adjustments.⁹⁵

In obese patients, increased body weight is related to decreased exposure, and dose adjustments based on weight have been suggested for anidulafungin, micafungin, and caspofungin.^96-98 However, these dosing recommendations have not been widely adopted.

Due to differences in PK parameters, monitoring drug levels could help optimize the dose and obtain the desired drug exposure. However, this requires multiple blood samples and can be challenging to implement. Further, the impact of therapeutic drug monitoring on outcomes like survival and microbiological clearance needs to be clarified.⁹⁵

4. Intra-abdominal candidiasis

For IAC, the recommended empiric treatment is also echinocandins, and a switch to fluconazole can be made using the same criteria as for candidemia.⁵⁹ Furthermore, source control and removing infected devices are essential for management, given the suboptimal penetration of echinocandins in peritoneal fluid, which might reduce their effectiveness in IAC.⁴⁷ The minimum treatment duration is 2 weeks, but duration depends on clinical response and source control.^47,59,60 Although the STOP-IT trial⁹⁹ demonstrated the safety of short antibiotic regimens in intra-abdominal infections (IAI), it excluded complex IAI, cases that are common in IAC^15,49. Also, Candida represented just 7% of the isolates in the short-term arm. Therefore, these findings may not be generalizable to IAC.

5. New antifungals

Rezafungin is a new echinocandin that was FDA approved in 2023. It has the benefit of being dosed weekly and has activity against most Candida species, including C. auris.¹⁰⁰ In addition to echinocandins, newer antifungal classes are being explored for their potential in treating invasive candidiasis. For instance, ibrexafungerp is the first triterpenoid antifungal agent, and it is currently under evaluation in adults with a fungal infection not cured with standard antifungals.¹⁰⁰

Manogepix/fosmanogepix is a first-in-class agent that targets the fungal cell wall. In phase 2 trials, it has demonstrated safety and efficacy in microbiological clearance in the treatment of non-neutropenic patients with candidemia, including due to C. auris^101,102. Other antifungals under investigation, initially for vulvovaginal candidiasis, are encochleated amphotericin and oteseconazole.¹⁰⁰ For highly-resistant isolates, infectious diseases consultants should consider use of novel antifungals such as ibrexafungerp and fosmanogepix.

Future directions

Several areas in managing invasive candidiasis need to be clarified in further research. First, updates on epidemiology of invasive candidiasis in the ICU are required. Second, new prediction models potentially using ML could help to guide decisions regarding empiric antifungal therapy. Moreover, in the diagnosis of invasive candidiasis, more evidence is needed to support molecular techniques for earlier identification without waiting for positive blood cultures. Additionally, questions remain about the role of routine therapeutic drug monitoring and appropriate dose adjustments in critically ill patients.

Summary

Invasive Candida infections are common among ICU patients due to the high prevalence of important risk factors such as antibiotic use, indwelling vascular catheters, and immunocompromise. The emergence of non-albicans species with potential for antifungal resistance, including C. auris, is concerning. Given the high mortality rate associated with invasive Candida infections, further research on preventive strategies is needed. Early recognition of patients at high risk for candidiasis, along with a structured, step-by-step approach that ensures timely diagnostic testing, prompt initiation of antifungal therapy, and complete source control, may improve survival and reduce the risk of complications.

Topic importance:

Invasive candidiasis (IC) is a common complication in patients in intensive care units (ICUs). IC can be divided into bloodstream infections (candidemia) and deep-seated infections. Candida infections are associated with high mortality and longer hospital stays. Moreover, the emergence of C. auris and other species prone to cause ICU outbreaks highlights the importance of this pathogen.

Review Findings:

Candidemia represents about 20% of bloodstream infections in ICU patients, while intra-abdominal candidiasis (IAC), the most common form of deep-seated infection, predominantly affects surgical ICU patients. Regarding species distribution, C. albicans is the most common species, although non-albicans species have become more prevalent in the past decade. Clinical signs of IC are indistinguishable from bacterial infections; therefore, identifying patients at risk of IC is key to choosing the right diagnostic approach. Diagnostic tests include conventional methods such as blood and tissue cultures; however, they have suboptimal sensitivity and are time-dependent, which delays the confirmation of IC. As a result, newer molecular tests offer faster results, though their use remains limited by cost and lack of external validation. Once IC is diagnosed first line treatment is with echinocandin antifungals, with the option of switching to azoles in selected patients. In addition, new antifungals are being tested to treat multidrug-resistant Candida species.

Summary:

Candida is a significant pathogen in ICU patients and has a heterogeneous clinical presentation. Conventional diagnostic methods lack sufficient sensitivity to detect IC at an early stage. New diagnostic strategies may support timely antifungal therapy.

Abbreviations list:

IC

invasive candidiasis

ICUs

intensive care units

IAC

intra-abdominal candidiasis

US

United States

WHO

World Health Organization

GIT

gastrointestinal tract

GUT

gastrointestinal-induced transition cell type

IFN-γ

interferon-gamma

PPV

positive predictive value

PCR

polymerase chain reaction

NPV

negative predictive value

ML

machine learning

CVC

central venous catheter

IDSA

Infectious Diseases Society of America

ESICM/ESCMID

European Society of Intensive Care Medicine and European Society of Clinical Microbiology and Infectious Diseases

CT

computed tomography

PK

pharmacokinetic

TDL

therapeutic drug levels

AUC

area under the curve

MIC

minimum inhibitory concentration

IAI

intra-abdominal infections