Management of candidemia and invasive candidiasis in adults

INTRODUCTION — Candidemia refers to presence of Candida species in the blood. Candida in a blood culture should never be viewed as a contaminant and should prompt evaluation for metastatic infection [1].

Invasive candidiasis refers to systemic Candida infection, in the presence or absence of candidemia; examples include osteoarticular infection and hepatosplenic candidiasis. Candidemia is the most common manifestation of invasive candidiasis.

The treatment of systemic candidal infection in adults will be reviewed here. Antifungal susceptibility testing and the treatment of candidemia in neonates and children are discussed separately. (See "Antifungal susceptibility testing" and "Treatment of Candida infection in neonates" and "Candidemia and invasive candidiasis in children: Management".)

The epidemiology, pathogenesis, clinical manifestations, and diagnosis of candidemia are discussed separately. (See "Overview of Candida infections" and "Candidemia in adults: Epidemiology, microbiology, and pathogenesis" and "Clinical manifestations and diagnosis of candidemia and invasive candidiasis in adults".)

Forms of invasive candidiasis are described separately. (See "Candida endocarditis and suppurative thrombophlebitis" and "Chronic disseminated candidiasis (hepatosplenic candidiasis)" and "Candida osteoarticular infections" and "Candida infections of the central nervous system".)

CLINICAL EVALUATION — Physicians caring for patients with candidemia should be aware that metastatic foci of infection can occur involving the eye (endophthalmitis), heart (endocarditis), kidney (pyelonephritis), joints, or other sites. The clinical manifestations of each condition are discussed in detail separately. (See "Overview of Candida infections" and "Clinical manifestations and diagnosis of candidemia and invasive candidiasis in adults" and "Epidemiology, clinical manifestations, and diagnosis of fungal endophthalmitis" and "Candida endocarditis and suppurative thrombophlebitis" and "Chronic disseminated candidiasis (hepatosplenic candidiasis)" and "Candida osteoarticular infections".)

●Role of ophthalmologic examination – We pursue routine ophthalmologic examination for all patients with candidemia to evaluate for endophthalmitis. In neutropenic patients, ophthalmologic examination should be deferred until the first week after neutrophil recovery because findings of choroidal and vitreal infection are minimal in the context of neutropenia. (See "Epidemiology, clinical manifestations, and diagnosis of fungal endophthalmitis", section on 'Clinical manifestations'.)

We believe the benefits of early diagnosis and appropriate management of Candida endophthalmitis (including administration of a prolonged course of an antifungal agent that achieves adequate intraocular concentration) can be vision-saving. Our approach is in alignment with the Infectious Diseases Society of America (IDSA) Candida guideline panel, which favors routine ophthalmologic examination for all patients with candidemia [1]. (See "Treatment of endogenous endophthalmitis and chorioretinitis due to Candida species".)

There is debate among professional groups as to the merit of this approach; in July 2021, the American Academy of Ophthalmology issued a statement favoring routine ophthalmologic examination only for patients with candidemia and either ocular symptoms or inability to report symptoms [2]. The statement cites a systematic review including 38 studies and more than 7400 patients with candidemia (including adults, children, and neonates) who underwent ophthalmologic evaluation. Some of the studies were designed to assess the incidence of eye involvement on routine ophthalmologic examination among patients with candidemia, while others were trials designed to compare various antifungal agents for treatment of candidemia. Endophthalmitis was observed in <1 percent of the cases, but the frequency varied markedly among the various studies. The authors voiced concern that endophthalmitis was over-reported in many of these studies when routine examinations were performed, and that this led to invasive interventions and associated harm [3].

While the incidence of endophthalmitis in patients with candidemia may be low, at least one study suggests that the incidence may be increasing, possibly due to the increased use of echinocandins to treat candidemia (echinocandins do not achieve significant intraocular concentrations) [4]. (See "Epidemiology, clinical manifestations, and diagnosis of fungal endophthalmitis", section on 'Endogenous Candida endophthalmitis'.)

●Role of echocardiography – Infective endocarditis (IE) should be suspected in non-neutropenic patients with persistently positive blood cultures (with or without physical stigmata of IE), especially in the setting of predisposing conditions such as prior IE or injection drug use (IDU) [5]. In such cases, echocardiography should be pursued. In one study of 20 patients with Candida IE, 65 percent had a history of IDU and 55 percent had a prior episode of bacterial IE [6]. (See "Candida endocarditis and suppurative thrombophlebitis".)

Routine echocardiography is not warranted for patients with candidemia in the absence of the above factors, given the rarity of this complication; in one study that included 187 adults with candidemia who underwent routine echocardiography, IE was diagnosed in 5.9 percent of cases [7]. (See "Candida endocarditis and suppurative thrombophlebitis".)

Among neutropenic patients, IE is unusual; such patients typically have positive cultures from a gastrointestinal source and are generally at low risk for endocarditis in the context of relative thrombocytopenia.

●Role of abdominal imaging – Uncommonly, liver and spleen abscesses can occur as a complication of candidemia and should be considered in people with abdominal symptoms, liver enzyme abnormalities, and/or persistent fever; in such cases, computed tomography of the abdomen should be pursued.

Hepatosplenic candidiasis (chronic disseminated candidiasis) is a distinct syndrome that is typically recognized after neutrophil recovery and bone marrow engraftment in previously neutropenic patients. This syndrome generally occurs in the context of a breakdown of gastrointestinal tract integrity with associated portal invasion, and usually manifests as fever of unknown etiology without candidemia. (See "Chronic disseminated candidiasis (hepatosplenic candidiasis)".)

MANAGEMENT APPROACH

Overview — Management of candidemia consists of early and appropriate antifungal therapy and targeted source control, along with individualized decisions regarding central venous catheter removal (if present) [8]. Catheter removal in the absence of concomitant antifungal therapy is not sufficient for management of candidemia [1]. (See 'Initial therapy' below and 'Central venous catheter removal' below.)

Blood cultures should be performed daily or every other day after initiation of antifungal therapy and catheter removal to establish clearance of candidemia. If blood cultures remain positive for several days after initiation of antifungal therapy and catheter removal, then a search for a metastatic focus, such as infective endocarditis or an abscess should be undertaken. (See 'Clinical evaluation' above.)

Initial therapy — Treatment of candidemia consists of prompt initiation of antifungal therapy [1,9-11]. Antifungal agents include the echinocandins, the azoles, and amphotericin B formulations. (See 'Antifungal agents' below.)

Susceptibility testing for azoles and echinocandins is warranted for all bloodstream and other clinically significant Candida isolates [1]. Specific drug resistance information for the various Candida species is discussed below. Antifungal susceptibility testing is discussed in greater detail separately. (See "Antifungal susceptibility testing".)

In general, treatment consists of monotherapy; no benefit for combination therapy has been established. In one trial, which included 219 non-neutropenic patients with candidemia and randomly assigned to treatment with either fluconazole (800 mg/day) monotherapy for two weeks or fluconazole (800 mg/day) plus amphotericin B (0.7 mg/kg per day) for the first four to seven days followed by fluconazole monotherapy to complete a two-week course of therapy after the blood cultures had cleared, the overall treatment success rates were similar [12].

Non-neutropenic patients — In non-neutropenic patients with candidemia, we suggest initial therapy with one of the following echinocandins (table 1) [1]. Dosing is as follows:

●Anidulafungin: 200 mg loading dose, then 100 mg intravenously (IV) daily

●Caspofungin: 70 mg loading dose, then 50 mg IV daily

●Micafungin: 100 mg IV daily

In non-neutropenic patients who are not critically ill and who do not have a fluconazole-resistant organism (such as Candida glabrata or Candida krusei), fluconazole (800 mg [12 mg/kg] loading dose on day 1, followed by 400 mg [6 mg/kg] orally or IV daily) can be used as an alternative agent.

Rezafungin, a newer echinocandin that can be administered as a once-weekly infusion (400 mg on day 1, then 200 mg once weekly beginning on day 8 for up to four doses), is an alternative. However, we favor other options that have more published clinical efficacy data and clinical experience [13,14]. The US Food and Drug Administration (FDA) has approved rezafungin for use only in adults 18 years and older who have candidemia and invasive candidiasis and who have limited or no alternative options [15]. Because of its long half-life, rezafungin may be considered in situations in which the patient is medically stable and ready for discharge but has an azole-resistant Candida spp (ie, C. glabrata or C. krusei) for which there is no oral therapy.

If there is intolerance, limited availability, or resistance to other antifungal agents, lipid formulations of amphotericin B (3 to 5 mg/kg IV daily) are an alternative.

This approach is supported by several randomized trials demonstrating that the echinocandins are as effective as and better tolerated than an amphotericin B formulation or fluconazole [16-19]. In one randomized trial including more than 500 patients with invasive candidiasis treated with either micafungin or liposomal amphotericin B (of whom 12 percent were neutropenic), success rates for clinical and microbiologic cure in a modified intent-to-treat analysis were similar (89.6 and 89.5 percent, respectively; difference 0.7 percent, 95% CI -5.3 to 6.7); renal dysfunction was more frequent among those treated with liposomal amphotericin B than micafungin (29.9 versus 10.3 percent), as were adverse events leading to treatment discontinuation (9 versus 4.9 percent) [17]. In a randomized trial that included 245 patients with candidemia who were treated with either anidulafungin or fluconazole (of whom 3 percent were neutropenic), the treatment success rate was higher among those treated with anidulafungin compared with fluconazole (75.6 versus 60.2 percent; difference 15.4, 95% CI 3.9-27.0) [18].

In addition, several randomized trials have shown that fluconazole is as effective as amphotericin B for treatment of candidemia [20-23]. In a meta-analysis including 15 randomized trials comparing antifungal agents for treatment of invasive candidiasis, there was no difference in mortality between fluconazole and amphotericin B (30 versus 33 percent; relative risk [RR] 0.92; 95% CI, 0.72-1.17); however, there was a higher rate of microbiologic failure in those patients who received fluconazole when compared with amphotericin B (27.4 versus 17.8 percent; RR 1.52, 95% CI 1.12-2.07) [24].

Neutropenic patients — In neutropenic patients with candidemia, we recommend initial therapy with one of the following echinocandins: micafungin, anidulafungin, or caspofungin; dosing is as summarized in the table (table 1) [1]. An alternative is a lipid formulation amphotericin B (3 to 5 mg/kg daily), but this regimen has more toxicity and appears to be less effective (table 1).

Fluconazole or other azoles should not be used as initial therapy in neutropenic patients who require antifungal therapy, given the widespread use of azole prophylaxis in neutropenic patients and the resulting increased prevalence of non-albicans Candida species with reduced susceptibility to fluconazole and voriconazole. Use of fluconazole should be reserved for patients who cannot be treated with other antifungal agents and whose absolute neutrophil count is above 500 cells/microL and rising. In one retrospective study, fluconazole monotherapy for C. glabrata fungemia was associated with worse survival than therapy with an echinocandin or amphotericin B [25].

No randomized trials have been adequately powered to evaluate the efficacy of antifungal therapy among neutropenic patients with candidemia [1]; available data are derived from small subset analyses of randomized trials, open-label studies, and retrospective studies [16,17,22,26-28]. Based on available data in non-neutropenic patients, the response rates to the echinocandins appear to be similar to or better than response rates to amphotericin B [18]. (See 'Non-neutropenic patients' above.)

Targeted and step-down therapy

General principles — In general, non-neutropenic and neutropenic patients who are clinically stable, who have Candida isolates that are susceptible to fluconazole, and who have negative repeat blood cultures can be transitioned to oral fluconazole five to seven days after the antifungal initiation, assuming the gastrointestinal tract is functional [1]. For most Candida species, fluconazole should be given at a dose of 400 mg (6 mg/kg) orally once daily for step-down therapy [1].

The approach to targeted therapy of candidemia due to fluconazole-resistant species is discussed below.

C. albicans — The approach to targeted and step-down therapy for patients with Candida albicans infection is as described above. (See 'General principles' above.)

The incidence of antifungal resistance among C. albicans remains low. Among 90,000 isolates of C. albicans collected from 40 countries between 1997 and 2005, only 1.5 percent were resistant to fluconazole [29]. More recent studies show resistance to fluconazole ranges between 0.3 and 2 percent [30-32].

Individual cases and small series of nonmucosal infections with fluconazole-resistant C. albicans have been reported from tertiary care centers; these usually occur in immunosuppressed patients on chronic fluconazole prophylaxis [33-35].

Most C. albicans isolates are susceptible to the echinocandins, although resistance has rarely been reported [36,37]. The vast majority of C. albicans isolates are susceptible to amphotericin B.

C. auris — In patients with infections due to Candida auris, we suggest targeted initial treatment with an echinocandin; dosing is summarized in the table (table 1) [38,39]. Because C. auris can develop resistance quickly, patients receiving antifungal therapy should be monitored carefully with follow-up surveillance blood cultures [38,39]. If the patient does not respond clinically to an echinocandin or has persistent candidemia for several days, treatment can be switched to a lipid formulation of amphotericin B 5 mg/kg IV daily. Azoles are generally not useful in the management of C. auris infections. Moreover, antifungal susceptibility testing should be done on all C. auris strains and used to guide individual selection of antifungal therapy [40].

C. auris is considered multidrug resistant with intrinsic resistance to agents from several antifungal classes, particularly azoles. Voriconazole resistance has been variable (3 to 73 percent), whereas posaconazole, itraconazole, and isavuconazole have displayed better activity [41]. Resistance to amphotericin B has been reported in 13 to 35 percent of isolates. Most C. auris isolates have been found susceptible to all of the echinocandins; however, minimum inhibitory concentrations (MICs) are higher than the MICs seen in C. albicans. Some isolates have had elevated MICs for all three major classes of antifungal agents (azoles, polyenes, and echinocandins) [42-44].

Susceptibility breakpoints for C. auris have not been established; however, based on susceptibility breakpoints established for other Candida species, most C. auris isolates have been highly resistant to fluconazole [39,41,42,44-47]. Tentative MIC resistance breakpoints are ≥32 mcg/mL for fluconazole, ≥2 mcg/mL for amphotericin B, ≥2 mcg/mL for caspofungin, and ≥4 mcg/mL for anidulafungin and micafungin [39,48]. These breakpoints are based on closely related Candida species and expert opinion (particularly for amphotericin B, for which there are no breakpoints for any Candida species).

Given concerns about resistance and transmission of C. auris in health care facilities, there are special screening recommendations and infection control precautions for patients who are colonized or infected with C. auris. (See "Candidemia in adults: Epidemiology, microbiology, and pathogenesis", section on 'C. auris emergence'.)

Additional details can be found on the United States Centers for Disease Control and Prevention website and Public Health England website.

C. dubliniensis — The approach to targeted and step-down therapy for patients with Candida dubliniensis infection is as described above. (See 'General principles' above.)

C. dubliniensis shares many phenotypic traits with C. albicans, and many isolates were previously misidentified as C. albicans. Special techniques must be undertaken in the microbiology laboratory to differentiate between these two species [49]. Most C. dubliniensis isolates are susceptible to azoles, echinocandins, and amphotericin B.

C. glabrata — For initial targeted treatment of candidemia due to C. glabrata, an echinocandin is preferred over amphotericin B. For patients with serious infection, we favor continuing treatment with an echinocandin.

Some studies have shown increased rates of echinocandin resistance among C. glabrata bloodstream isolates [50-53]. Resistance should be suspected in patients who have received echinocandins in the recent past and in patients who develop candidemia while receiving an echinocandin for prophylaxis or empiric therapy (eg, for neutropenic fever); in these situations, an amphotericin B formulation should be used until antifungal susceptibility testing results are available. Amphotericin B has delayed killing kinetics against C. glabrata in vitro [54]; therefore, higher doses of amphotericin B are recommended when treating known C. glabrata infection (1 mg/kg daily of amphotericin B deoxycholate or 5 mg/kg daily of lipid-based formulations).

Patients who were treated initially with an echinocandin or amphotericin B, are clinically stable, and have negative blood cultures can be transitioned to oral fluconazole after five to seven days of antifungal therapy. In order to step-down to an azole and to oral therapy, in vitro susceptibility data are required. Management of infections caused by C. glabrata isolates with MICs that are considered to be “susceptible” or “susceptible dose dependent” (implying that higher doses of fluconazole should be used [fluconazole 800 mg (12 mg/kg orally once daily)]) or voriconazole 200 to 300 mg (3 to 4 mg/kg orally twice daily) should be given. In a retrospective study including 127 patients with candidemia due to C. glabrata, higher doses of fluconazole (average dose ≥400 mg) were found to be an independent predictor of survival (odds ratio [OR] 3.96, 95% CI 1.52-10.42) [55]. While these data do not address step-down therapy, they support use of higher fluconazole doses for treatment of susceptible C. glabrata.

Fluconazole MIC thresholds are higher for C. glabrata than for many other Candida species. The Clinical and Laboratory Standards Institute designates C. glabrata isolates with MICs of 32 mcg/mL as "susceptible dose dependent" but many clinicians would not use fluconazole for step-down treatment of C. glabrata infections with an MIC this high [56]. In such cases, an echinocandin (or amphotericin B) would be continued for the duration of therapy.

Many C. glabrata isolates are now intrinsically resistant to the azoles, mostly due to changes in drug efflux [1,57]. This type of resistance can sometimes be overcome by using higher doses of fluconazole, but in general, this approach is not recommended for candidemias or invasive infections. Cross-resistance between fluconazole and voriconazole is also common with C. glabrata. C. glabrata is known to exhibit phenotypic heteroresistance to fluconazole, which has been associated with persistent infection [58]. Furthermore, among the Candida species, voriconazole MICs are highest with C. glabrata. Fluconazole-resistant isolates are generally voriconazole-resistant as well [50,59]. (See "Overview of antibacterial susceptibility testing", section on 'Heteroresistance'.)

There is concern that some C. glabrata bloodstream isolates with resistance to fluconazole and voriconazole can also become resistant to the echinocandins [50,60-62]. In a study including 1669 C. glabrata bloodstream isolates collected in the United States between 2006 and 2010, 162 isolates (9.7 percent) were resistant to fluconazole, of which 98.8 percent were also not susceptible to voriconazole. In addition, 9.3, 9.3, and 8.0 percent of the isolates were also found to be nonsusceptible to anidulafungin, caspofungin, and micafungin, respectively [50]. Of the 162 isolates that were resistant to fluconazole, 18 (11.1 percent) were also resistant to one or more of the echinocandins, with associated mutations in FKS1 or FKS2 genes. In comparison, there were no echinocandin-resistant strains detected among 110 fluconazole-resistant C. glabrata isolates tested between 2001 and 2004. This is noteworthy because echinocandins were used sparingly in that time period. Similarly, in a study including 1380 C. glabrata bloodstream isolates collected in the United States between 2008 and 2013, 3 to 4 percent of strains were resistant to all three echinocandins, and approximately one-third of echinocandin-resistant strains were resistant to fluconazole [60]. Nearly all of the isolates with an FKS1 or FKS2 mutation were resistant to at least one echinocandin.

C. guilliermondii — Targeted therapy for patients with Candida guilliermondii should be guided by susceptibility testing. An echinocandin can be used as initial therapy while awaiting susceptibility testing.

Some C. guilliermondii isolates have reduced susceptibility to fluconazole and others have reduced susceptibility to echinocandins [63,64]. C. guilliermondii isolates may be susceptible to other azoles, such as voriconazole and posaconazole, but often fluconazole-resistant strains are cross-resistant to other azoles as well. However, C. guilliermondii is usually susceptible to amphotericin B.

C. krusei — For targeted treatment of candidemia due to C. krusei, an echinocandin is the preferred antifungal agent. In general, most C. krusei isolates are resistant to fluconazole but remain susceptible to the echinocandins [65]; however, rare cases of echinocandin resistance have been reported [66-68]. C. krusei isolates can also demonstrate decreased susceptibility to amphotericin B; for this reason, higher doses of amphotericin B are warranted for the treatment of C. krusei infections (5 mg/kg daily of lipid-based formulations).

For oral step-down therapy of C. krusei infection, voriconazole is recommended; C. krusei is intrinsically resistant to fluconazole due to an altered cytochrome P450 isoenzyme; this resistance cannot be overcome with use of higher drug doses [69]. Voriconazole binds more effectively to the cytochrome P450 isoenzyme in C. krusei than fluconazole, resulting in higher susceptibility rates [70]. Cross-resistance between voriconazole and fluconazole does not occur with C. krusei, as has been described with C. glabrata.

There are geographic differences in the incidence of voriconazole-resistant C. krusei. In an international surveillance study, which included 326 bloodstream isolates of C. krusei, voriconazole resistance was observed in only 7.4 percent of isolates; such resistance was most common among isolates from Latin America and uncommon among isolates from North America and Europe [65].

C. lusitaniae — The approach to targeted and step-down therapy for patients with Candida lusitaniae infection is as described above. (See 'General principles' above.)

C. lusitaniae is unique among Candida species in that it is often resistant to or quickly becomes resistant to polyenes, specifically amphotericin B; however, it is usually susceptible to the azoles and echinocandins [71].

C. parapsilosis — The approach to targeted and step-down therapy for patients with Candida parapsilosis infection is as described above. (See 'General principles' above.)

C. parapsilosis is highly susceptible to most antifungal agents. In a surveillance study including 9371 C. parapsilosis isolates collected between 2001 and 2005, high rates of fluconazole susceptibility (91 to 96 percent) and voriconazole susceptibility (95 to 98 percent) were observed in all geographic regions except Africa and the Middle East (79 and 86 percent susceptible to fluconazole and voriconazole, respectively) [72]. Subsequent surveillance studies have demonstrated fluconazole resistance rates ranging from 2 to 6 percent [30-32].

C. parapsilosis MICs for echinocandins are higher than for other Candida species; however, the clinical implications of these in vitro susceptibility data are uncertain [57,73]. In one review of the data from five different treatment trials of caspofungin use in patients with invasive candidiasis, the overall (clinical and microbiologic) success rate among patients with C. parapsilosis (74 percent) was similar to patients with invasive candidiasis caused by other Candida species [74]. A multicenter, prospective observational study of candidemia due to C. parapsilosis in Spain found no differences in outcomes of patients who were treated with an echinocandin compared with those who were treated with an azole [75]. These results should be interpreted with caution because this was not a randomized trial; however, it is unlikely that such a study will be performed [76].

C. tropicalis — The approach to targeted and step-down therapy for patients with Candida tropicalis infection is as described above. (See 'General principles' above.)

C. tropicalis is usually susceptible to the azoles, amphotericin B, and the echinocandins. However, breakthrough C. tropicalis bloodstream infections with resistance to caspofungin have been reported rarely in patients with hematologic malignancies [37,77,78].

Duration of therapy — The optimal duration of therapy for candidemia is uncertain.

For patients with candidemia in the absence of metastatic complications, a minimum of two weeks of therapy after blood cultures become negative has been used in most clinical trials and is the recommended duration in the 2016 Infectious Diseases Society of America (IDSA) guidelines [1]. In addition, patients should have resolution of symptoms attributable to candidemia and resolution of neutropenia (eg, absolute neutrophil count >500 cells/microL and a consistent increasing trend) prior to discontinuation of antifungal therapy [1]. (See "Overview of neutropenic fever syndromes", section on 'Neutropenia' and "Treatment of neutropenic fever syndromes in adults with hematologic malignancies and hematopoietic cell transplant recipients (high-risk patients)", section on 'Duration'.)

For patients with candidemia and metastatic foci of infection (such as endophthalmitis or endocarditis), a longer duration of therapy and consultation with an infectious disease specialist are warranted. Issues related to the management of these conditions are discussed further separately. (See "Treatment of endogenous endophthalmitis and chorioretinitis due to Candida species" and "Candida endocarditis and suppurative thrombophlebitis" and "Candida osteoarticular infections" and "Chronic disseminated candidiasis (hepatosplenic candidiasis)".)

Central venous catheter removal — We are in agreement with the 2016 IDSA guidelines regarding central venous catheter (CVC) removal in the setting of candidemia [1]:

●For non-neutropenic patients with candidemia whose source is presumed to be a CVC, the CVC should be removed as early as possible. Of course, the timeframe should be individualized for each patient.

●For neutropenic patients with candidemia, the risk-benefit ratio of CVC removal is less straightforward. Management decisions should be individualized with a careful consideration of the likely source of candidemia (eg, catheter versus gastrointestinal source) and the risk of catheter removal (eg, bleeding):

•Among patients with neutropenia in the context of hematologic malignancy, chemotherapy-induced mucositis is common and likely to serve as a gastrointestinal source of candidemia; in such patients, routine catheter removal may not be necessary. In this setting, risks of removal include thrombocytopenia-associated bleeding and risks associated with loss of intravenous access [1,79-82].

•Among patients with neutropenia in the absence of gastrointestinal mucositis, the likelihood of catheter-related infection is greater and catheter removal is more important.

All patients with candidemia must be treated with antifungal therapy; catheter removal in the absence of concomitant antifungal therapy is not sufficient for the management of candidemia [1]. (See 'Initial therapy' above and 'Targeted and step-down therapy' above.)

No randomized trials have assessed the efficacy of CVC removal on the resolution of candidemia [83]. However, limited data suggest that candidemia clears more quickly when CVCs are removed [84,85], and leaving CVCs in place may be associated with increased mortality [8,75,85-88]:

●In a quantitative review of data from seven randomized trials of candidemia and invasive candidiasis (mostly in non-neutropenic patients), CVC removal was associated with decreased mortality (OR 0.50, 95% CI 0.35-0.72) [8].

●In a retrospective subgroup analysis of two randomized trials evaluating antifungal therapy in patients with candidemia, early CVC removal (within 24 or 48 hours) did not improve time to candidemia clearance or rates of persistent or recurrent candidemia, but was associated with improved treatment success and survival; however, in multivariate analysis these benefits were not observed [80].

●In a study of patients with cancer, a fivefold difference in quantitative cultures taken from the CVC and a peripheral vein was used to define catheter-associated candidemia; using these criteria, patients with noncatheter sources did not benefit from CVC removal [89].

There are multiple limitations to observational studies and subgroup analyses of randomized trials, including unrecognized confounders, treatment bias, lack of standardized criteria for catheter removal or data on time to removal, and lack of statistical power [90,91].

ROLE OF EMPIRIC ANTIFUNGAL THERAPY — In the absence of bloodstream infection or evidence of invasive candidiasis based on deep tissue culture, the approach to empiric antifungal therapy depends on individual clinical circumstances:

●For patients with neutropenic fever, empiric antifungal therapy is administered routinely given the substantial risk for developing invasive candidiasis. This is discussed in detail separately. (See "Treatment of neutropenic fever syndromes in adults with hematologic malignancies and hematopoietic cell transplant recipients (high-risk patients)", section on 'Addition of an antifungal agent'.)

●For non-neutropenic patients with sepsis, routine administration of antifungal therapy as part of initial management is not routinely warranted. This is discussed further separately. (See "Evaluation and management of suspected sepsis and septic shock in adults", section on 'Choosing a regimen'.)

●For non-neutropenic, critically ill patients with persistent fever or unexplained hypotension despite broad-spectrum antibacterial therapy, the benefit of empiric antifungal therapy is uncertain. Clinical criteria to guide antifungal initiation decisions regarding its use remain poorly defined.

In general, we are in agreement with the 2016 Infectious Diseases Society of America (IDSA) guidelines, which state that empiric antifungal therapy should be considered in critically ill patients with persistent, unexplained fever and risk factors for invasive candidiasis [1]. The relevant risk factors include presence of an indwelling central venous catheter, administration of parenteral nutrition, need for hemodialysis, trauma or burns, use of broad-spectrum antibiotics, recent surgery (particularly abdominal surgery), positive serum beta-D-glucan assay, and cultures demonstrating Candida colonization at nonsterile sites. (See "Candidemia in adults: Epidemiology, microbiology, and pathogenesis".)

In such cases, we favor the use of an echinocandin. Fluconazole is an acceptable alternative agent in patients who have not received prior azoles and are known to be colonized with fluconazole-susceptible Candida species.

For patients who receive empiric antifungal therapy but do not respond after four to five days and have no evidence of invasive candidiasis (including a negative serum beta-D-glucan assay, which has high negative predictive value), discontinuation of empiric antifungal therapy is reasonable [1].

Thus far, clinical studies evaluating the role of empiric antifungal therapy for non-neutropenic patients with ongoing signs of infection in the setting of broad-spectrum antimicrobial therapy have not demonstrated substantial benefit. In a randomized trial including 260 non-neutropenic critically ill patients with Candida colonization, multiorgan failure, and intensive care unit-acquired sepsis, empiric treatment with micafungin for 14 days did not result in improved infection-free survival at 28 days. However, it did reduce the rate of new fungal infections [92]. Similarly, in a randomized trial including 241 critically ill patients who underwent surgery for intra-abdominal infection and were treated with either micafungin or placebo, no difference in the incidence of invasive candidiasis was observed (11.1 versus 8.9 percent; difference 2.24 percent, 95% CI -5.52 to 10.20) [93]. Some earlier studies in patients who underwent bowel surgery or had bowel perforation demonstrated some benefit of empiric fluconazole in this setting [94-96].

ANTIFUNGAL AGENTS — The most common antifungal agents used for the treatment of candidemia are echinocandins (caspofungin, micafungin, anidulafungin; rezafungin is not commonly used) and fluconazole; these agents are better tolerated than amphotericin B formulations, which are used less often due to their risk of toxicity [24].

Echinocandins — The echinocandins include caspofungin, anidulafungin, micafungin, and rezafungin. Echinocandins are noncompetitive inhibitors of the synthesis of 1,3-beta-D-glucan, which is an integral component of the fungal cell wall [97]. Echinocandins are fungicidal for most Candida species, have favorable toxicity profiles, and are approved for the treatment of candidemia and other forms of invasive candidiasis. Dosing is summarized in the table (table 1). Adverse effects of echinocandins are generally mild and include fever, thrombophlebitis, headache, and elevated aminotransferases [98]. The pharmacology of the echinocandins is discussed in detail separately. (See "Pharmacology of echinocandins and other glucan synthesis inhibitors".)

The echinocandins are used extensively for candidemia and invasive candidiasis given their broad-spectrum activity against Candida species. The echinocandins are preferred over azoles for the initial treatment of candidemia and invasive candidiasis in the 2016 Infectious Diseases Society of America (IDSA) guidelines for management of candidiasis [1].

Echinocandin resistance had been observed in only a few individual cases; however, acquired resistance has been increasingly reported, especially in C. glabrata. The highest echinocandin minimum inhibitory concentrations (MICs) are for C. parapsilosis and C. guilliermondii. The mechanism of echinocandin resistance involves mutations in the FKS1 or FKS2 genes that control the enzyme targeted by the echinocandins [61,62,73,99]. If an isolate demonstrates echinocandin resistance with acquired mutations in FKS gene “hot spots,” it is typically resistant to the entire class; however, exceptions have been reported, specifically in Candida kefyr isolates [100].

Azoles — Available azoles include fluconazole, voriconazole, posaconazole, itraconazole, and isavuconazole. The azoles inhibit the cytochrome P450-dependent enzyme lanosterol 14-alpha-demethylase [101]. This enzyme is necessary for the conversion of lanosterol to ergosterol, a vital component of the fungal cellular membrane. Dosing is summarized in the table (table 1). The pharmacology of the azoles is discussed separately.

Azoles interact with multiple different cytochrome P450 enzymes; alternative antifungal agents, such as echinocandins, may be preferred if patients are taking other medications that utilize P450 pathways. Details about specific interactions may be obtained by using the drug interactions program included within UpToDate.

Fluconazole has an excellent safety profile, is available in intravenous and oral formulations, and is inexpensive because it is generic. Fluconazole is highly bioavailable, making oral dosing appropriate for most patients [12,20-22].

Other available azoles include voriconazole, posaconazole, and isavuconazole.

●The activity of voriconazole against Candida species is superior to that of fluconazole, with MICs that are a log or more lower than fluconazole [102]. However, cross-resistance between fluconazole and voriconazole has been observed, especially with C. glabrata. Voriconazole has significantly greater in vitro activity against C. krusei isolates compared with fluconazole [103]. In non-neutropenic patients, voriconazole was shown to be noninferior to amphotericin B for treatment of candidemia and was associated with fewer toxicities [104]. However, marginal benefits over fluconazole and increased relative toxicities limit its utility for primary therapy.

●Posaconazole is approved for oropharyngeal candidiasis but not for systemic candidiasis. It is also approved for use as a prophylactic agent for fungal infections in allogeneic hematopoietic cell transplant recipients with graft-versus-host disease and in patients with prolonged neutropenia due to chemotherapy for hematologic malignancies. Posaconazole has not been studied extensively for candidemia and may be considered to have marginal benefits over fluconazole.

●Isavuconazole is not approved for the treatment of candidiasis. It is formulated as a prodrug, isavuconazonium, which is quickly broken down to the active agent, isavuconazole. In a randomized trial that included 450 non-neutropenic patients, the endpoint of noninferiority to caspofungin for candidemia was not reached; caspofungin had a higher success rate [105]. Isavuconazole would be predicted to have only marginal benefits over fluconazole.

Amphotericin B — Amphotericin B is a polyene antifungal agent that disrupts fungal cell wall synthesis because of its ability to bind to sterols, primarily ergosterol, which leads to the formation of pores that allow leakage of cellular components. Dosing is summarized in the table (table 1). Issues related to pharmacology are discussed separately.

Amphotericin B deoxycholate demonstrates rapidly fungicidal in vitro activity against most species of Candida; it was the standard drug for the treatment of candidiasis for decades. It is rarely used anymore given its association with significant nephrotoxicity and its associated infusion-related adverse events. Lipid-based amphotericin B formulations include liposomal amphotericin B and amphotericin B lipid complex and are associated with less toxicity than amphotericin B deoxycholate.

Amphotericin B formulations are often avoided due to greater toxicity compared with azoles and echinocandins. They remain useful in cases when resistance to the other antifungal classes is suspected or proven.

OUTCOMES — Untreated, candidemia has a mortality rate of over 60 percent [106]. With treatment, the mortality is approximately 25 to 40 percent [8,91,107]. Treatment delay has been associated with increased mortality [10,11,108]. In a retrospective study including 230 patients with candidemia, the number of days between identification of candidemia and initiation of fluconazole correlated with increased mortality: zero days (15 percent), one day (24 percent), two days (37 percent), and three days (41 percent) [10].

Neutrophil recovery is tantamount to recovery from candidemia/candidiasis; mortality rates in neutropenic patients with candidemia approach 100 percent [109].

Among candidemic patients with septic shock, the outcomes are even poorer. In a retrospective study including 224 patients with septic shock and a positive blood culture for Candida spp, 64 percent died [9]. The in-hospital mortality rate among the 142 patients who had adequate source control within 24 hours of onset of septic shock and antifungal therapy administered within 24 hours of onset of septic shock was 53 percent; among the 82 patients for whom these goals were not attained, the mortality rate was 98 percent. On multivariate analysis, both failure to achieve timely source control (adjusted odds ratio [aOR] 77.4, 95% CI 21.5-278.4) and delayed antifungal therapy (aOR 33.8, 95% CI 9.7-118.0) were associated with a high mortality rate.

Factors associated with increased mortality in several other studies of hospitalized patients with candidemia include higher APACHE II scores, inadequate fluconazole dosing, infection with an echinocandin-resistant strain of C. glabrata, C. parapsilosis, or C. tropicalis, retention of a central venous catheter, increasing age, and use of immunosuppressive therapy [8,87,110,111]. In intensive care unit (ICU) patients, diabetes mellitus, immunosuppression, and mechanical ventilation were associated with death in one study [112]. Whereas in non-ICU patients, glucocorticoid use at the time of a positive blood culture was associated with increased mortality in a separate study [87].

Among cancer patients, persistent neutropenia, higher APACHE III score, and visceral dissemination have all been associated with poor prognosis [81].

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Candidiasis".)

SUMMARY AND RECOMMENDATIONS

●Definitions – Candidemia refers to presence of Candida species in the blood. Invasive candidiasis refers to systemic Candida infection in the presence or absence of candidemia. Candidemia is the most common manifestation of invasive candidiasis. (See 'Introduction' above.)

●Organ involvement in candidemia – Physicians caring for patients with candidemia should be aware that metastatic foci of infection can occur involving the eye (endophthalmitis), heart (endocarditis), kidney (pyelonephritis), joints, or other sites. (See 'Clinical evaluation' above.)

●Clinical management of confirmed candidemia – Management consists of antifungal therapy and individualized decisions regarding central venous catheter removal (if present). (See 'Overview' above.)

•Initial antifungal therapy – Selection of antifungal agents depends on the presence or absence of neutropenia.

-Non-neutropenic patients – We suggest initial therapy with an echinocandin rather than fluconazole or amphotericin B (table 1) (Grade 2B). (See 'Non-neutropenic patients' above.)

-Neutropenic patients – We recommend initial therapy with an echinocandin rather than an azole or amphotericin (table 1) (Grade 1B). Because of increased risk of fluconazole resistance due to prior azole use, fluconazole and other azoles should not be used as initial therapy in neutropenic patients. (See 'Neutropenic patients' above.)

•Transition to oral therapy – In general, patients who are clinically stable, have fluconazole-susceptible isolates, and have negative repeat blood cultures can be transitioned to oral fluconazole after five to seven days. A targeted approach to step-down therapy by species is summarized above. (See 'Targeted and step-down therapy' above.)

•Role of ophthalmologic examination – We pursue formal ophthalmologic examination for all patients with candidemia to evaluate for endophthalmitis. (See 'Clinical evaluation' above.)

•Central venous catheter management – For non-neutropenic patients with candidemia, we suggest central venous catheter (CVC) removal as soon as possible (Grade 2C). For neutropenic patients with candidemia, catheter management should be individualized, with consideration of the likely source of candidemia (eg, catheter versus gastrointestinal source) and the possible risks of catheter removal (eg, bleeding). (See 'Central venous catheter removal' above.)

Catheter removal in the absence of concomitant antifungal therapy is never sufficient for the management of candidemia.

•Monitoring response – Blood cultures should be performed daily or every other day to establish clearance of candidemia. If blood cultures remain positive for several days after the initiation of appropriate antifungal therapy and catheter removal, a search for a metastatic focus, such as endocarditis or abscess, should be undertaken. Endocarditis should be suspected in non-neutropenic patients with persistently positive blood cultures, especially in the setting of predisposing conditions such as prior endocarditis or injection drug use. (See 'Clinical evaluation' above and 'Overview' above.)

•Duration of therapy – For non-neutropenic patients in the absence of metastatic complications, we treat for a minimum of two weeks after the blood cultures become negative. Neutropenic patients should have resolution of symptoms attributable to candidemia and resolution of neutropenia (eg, absolute neutrophil count >500 cells/microL and rising) prior to discontinuation of antifungal therapy. Patients with metastatic foci of infection (eg, eye, bone, heart, liver, spleen) require a longer duration of therapy. (See 'Duration of therapy' above.)

●Clinical management of suspected candidemia – Because blood cultures are often negative in patients with candidemia, empiric antifungal therapy is indicated in certain situations.

•Neutropenic patients – For patients with neutropenic fever, empiric antifungal therapy is administered routinely given substantial risk for invasive candidiasis. This is discussed in detail separately. (See 'Role of empiric antifungal therapy' above and "Treatment of neutropenic fever syndromes in adults with hematologic malignancies and hematopoietic cell transplant recipients (high-risk patients)", section on 'Addition of an antifungal agent'.)

•Non-neutropenic patients – For critically ill patients with persistent fever or unexplained hypotension despite broad-spectrum antibacterial therapy, the benefit of empiric antifungal therapy is uncertain and the decision should be individualized and based on clinical assessment, surrogate markers (eg, serum beta-D-glucan), and culture data demonstrating Candida colonization at nonsterile sites. (See 'Role of empiric antifungal therapy' above.)