Treatment of invasive candidiasis: between guidelines and daily clinical practice

Abstract

Invasive candidiasis, including candidemia (IC/C), is a major cause of morbidity and mortality among hospitalized patients. While incidence is higher in intensive care units, the majority of cases of candidemia are documented in medical wards. Although Candida albicans is still the most frequently isolated species, IC/C due to non-albicans species, usually less susceptible to fluconazole, is increasing. Early identification of patients at risk, knowledge of local epidemiology and prompt efforts to define etiologic diagnosis play a pivotal role for appropriateness. Starting therapy with an echinocandin, switching then to fluconazole when possible, seems to represent a potentially useful strategy for the management of IC/C. The choice between the three echinocandins should be based on the specific indications, pharmacokinetic/pharmacodynamic profile, clinical experience and relative cost.

Keywords:

• antifungals
• candidemia
• invasive candidiasis

Invasive candidiasis, including candidemia (IC/C), is a major cause of morbidity and mortality among patients undergoing major surgery, receiving broad-spectrum antibiotic therapy, having severe clinical conditions, multiple comorbidities and indwelling devices [1,2].

A recent multicenter study conducted in Italy and Spain found that the overall incidence of candidemia was 1.55 cases per 1000 admissions; the majority of the candidemia episodes were found in the internal medicine department (49.6%), followed by the surgical ward, the intensive care unit (ICU) and the hemato-oncology ward; 30-day crude mortality rate was about 40% [3].

Candida albicans is the most frequently isolated species in the case of candidemia, but infections due to non-albicans species are increasing, especially in ICU [4,5]; the most common non-albicans species are C. parapsilosis and C. glabrata [6], followed by C. tropicalis, C. krusei, C. guilliermondii and C. lusitaniae [7]. Significant geographical variations exist. For instance, among 2085 clinical Candida isolates obtained from 79 medical centers, between January 2008 and December 2009, C. glabrata isolates were more common in North America (23.5%), and C. albicans isolates were more common in Asian Pacific (56.9%), with C. parapsilosis (25.6%) and C. tropicalis (17.0%) being more prominent in Latin America [5].

C. glabrata has a dose-dependent susceptibility to fluconazole while C. krusei has an innate resistance to the azoles [8]. C. parapsilosis, often associated with central venous catheter (CVC) infections, is less susceptible in vitro to the echinocandins. Amphotericin B has a good antifungal activity against all isolates of C. albicans and virtually all the non-albicans species.

The outcome of candidemia seems to be closely related to Candida species: fungemia with a more favorable outcome are those caused by C. parapsilosis, followed by those due to C. albicans, C. glabrata, C. tropicalis and C. krusei [9].

For these reasons, the knowledge of local epidemiology plays a pivotal role in planning an appropriate empirical treatment for Candida infections.

Timely and appropriate antifungal therapy is crucial for patient outcome: any delay is associated with an increased rate of mortality [10].

Another independent variable associated with mortality is represented by the production of biofilm, the extracellular matrix produced by microorganisms under particular circumstances, typically in the presence of foreign bodies such as indwelling devices and prosthesis [11]. Azoles lack any significant activity against biofilm while Echinocandins show variable efficacy; amphotericin B shows consistent in vitro activity against biofilm produced by different Candida species [12]. Treating infections due to biofilm-forming Candida with highly active anti-biofilm antifungal agents like caspofungin seems to favorably influence patient survival with respect to fluconazole therapy [13].

For a timely antifungal treatment, an empirical, fever-driven approach is often used [14]. This strategy, although largely adopted in the setting of patients with hematologic malignancies, did not show consistent efficacy in ICU patients. Schuster et al. [15] compared high-dose intravenous fluconazole (800 mg) to placebo in persistently febrile ICU patients not responding to antibacterial therapy. Quite unexpectedly, no difference was documented between the two treatment groups. Two possible explanations for this result were represented by the intrinsic limit of fluconazole antifungal activity and by the patient population unselected for high risk of IC/C.

Several predictive algorithms and score systems have been proposed to identify patients at higher risk of candidemia and invasive candidiasis. The ‘Candida score’ by Leon [16] is based on total parenteral nutrition (one point), recent surgery (one point), multi-focal Candida colonization (one point) and severe sepsis (two points). With a score of 3 or greater, the relative risk of developing fungemia and invasive candidiasis is increasing and the start of empirical antifungal therapy may be justified. The Candida colonization index (CI) is defined as the ratio of the number of distinct non-blood body sites colonized by Candida spp. to the total number of body sites cultured [17]. Patients with a CI reaching the threshold of 0.5 have a higher probability of developing invasive candidiasis. The corrected colonization index is defined as the product of the CI and the ratio of the number of distinct body sites showing heavy growth by semiquantitative culture to the total of distinct body sites with Candida spp. growth. Patients with a corrected colonization index reaching the threshold of 0.4 have a higher probability of developing invasive candidiasis. Empirical treatment based on IC and IC/C has been proven to reduce the incidence of invasive candidiasis. However, intense laboratory work is required and a high proportion of patient is treated. Ostrosky-Zeichner et al. identified the following predictive rule for IC among patients in ICU: any systemic antibiotic (days 1–3) OR presence of a central venous catheter (days 1–3) AND at least TWO of the following – total parenteral nutrition (days 1–3), any dialysis (days 1–3), any major surgery (days 7–0), pancreatitis (days 7–0), any use of steroids (days 7–3) or use of other immunosuppressive agents (days 7–0) [18]. Overall, these score systems have by far a high negative predictive value but a low positive predictive value. Cost–effectiveness remains to be proven in large prospective clinical trials. Additionally, these score systems have been designed for the surgery and ICU setting and may be unfit for the medical wards.

Systematic efforts to obtain an etiologic diagnosis have to be taken to ensure an appropriate treatment. Culture methods still play a key role in the diagnosis of IC/C. However, sensitivity of blood culture in the case of candidemia is around 50%. Recently, Tascini et al. showed that cultures taken from arterial blood presented a shorter time to positivity with respect to blood cultures drawn from peripheral vein; it is noteworthy that the time sparing was around 12 h. However, no significant difference was documented in time to positivity for arterial blood and blood drawn from CVC [19].

Serological methods for the diagnosis of invasive candidiasis include the combined detection of mannan antigen and anti-mannan antibodies and the β-glucan antigen. Mannan is a genus specific antigen produced by Candida in the early stages of the infection, but a lythic enzyme clears it rapidly from the serum. Sensitivity may be increased by the concomitant search of the anti-mannan antibodies that become positive at a later stage but can be detected for a longer period [20].

The β-glucan is a panfungal test useful for the detection of Candida, Aspergillus and Pneumocystis jiroveci. The test has a low specificity due to several causes of false-positive results. Trend in the β-glucan levels seems to be useful in predicting the outcome of invasive candidiasis and the response to antifungal therapy [21].

Significant advances have also been made in nucleic acid amplification methods for rapid detection of Candida in blood specimens but further evaluation of these approaches in different clinical settings is needed [22].

Guidelines for the treatment of invasive candidiasis have been proposed by the Infectious Diseases Society of America (IDSA) in 2009 [8] and by the European Society of Clinical Microbiology and Infectious Diseases (ESCMID) [14] in 2012.

Both guidelines emphasize the role of echinocandins. Echinocandins are indicated as first-line treatment of IC/C because of fungicidal activity, activity against strains producing biofilms, activity against fluconazole-resistant strains, favorable safety profile and low propensity for drug–drug interactions.

The choice between the three echinocandins in real life is usually based on cost. However, slight peculiarities in terms of indications, pharmacokinetic/pharmacodynamic profile and clinical experience exist. Caspofungin is the most widely used echinocandin with a large amount of data related to the efficacy and safety profile; it has a lower distribution volume and a higher serum concentration compared to the other echinocandins, suggesting a possible advantage in the setting of bloodstream infections, which however needs to be demonstrated [23]. Anidulafungin lacks any metabolism (biodegradation) and may be useful in patients with severe liver disease. Micafungin, although limited by the EMA warning, is indicated for neonates.

Conventional amphotericin B is no more considered an option in the treatment of candidemia by the ECCMID guidelines due to toxicity. Instead, liposomal amphotericin B is considered as an alternative to echinocandins, for critically ill patients with candidemia. The wider spectrum of activity of liposomal amphotericin B is counterbalanced by a higher cost and a less favorable safety profile, compared to echinocandins, so that pros and cons should be thoroughly considered in each specific situation.

Fluconazole was considered an option as first-line treatment of candidemia in stable patients by the 2009 IDSA guidelines (AI), based on the evidence that less critical patients with IC/C treated with fluconazole or anidulafungin had a similar mortality [24]. Conversely, it has been severely downgraded by the ECCMID guidelines (CI), mainly on the basis of Schuster’s study [15]. Given its limited spectrum of activity and the lack of fungicidal activity and considering the availability of more effective drugs, the use of fluconazole in critical patients and empirical treatment of candidemia seems not to be justified anymore. However, fluconazole may probably still be considered in stable patients with IC and candidemia sustained by susceptible strains. De-escalation from an echinocandin to intravenous or oral fluconazole should be encouraged when the patient is clinically stable and the isolated strain is susceptible.

In the case of Candida endophthalmitis, voriconazole is the drug of choice. Some kind of ocular involvement has been reported to occur in 16% and endophthalmitis in 1.6% of cases of candidemia [25]. Since echinocandins do not adequately penetrate in the eye, ocular involvement should be ruled out in all patients with candidemia.

Patients should be treated for at least 14 days after the last positive blood culture and even more in deep-seated infections.

Intravascular non-surgical catheters should be removed in all patients with documented catheter-related fungemia.

Main characteristics of antifungals are summarized in Table 1.

Table 1. Summary of the main characteristics of antifungal agents for treating invasive candidiasis/candidemia.

	Fluconazole	Echinocandins	Amphotericin B
Spectrum of activity	Limited	Wide	Wide
Antifungal activity	Fungistatic	Usually fungicidal	Fungicidal
Anti-biofilm activity	Low	Variable	Consistent
Safety	Good	Very good	Nephrotoxicity
Evidence from RCTs	Inferior to echinocandins	Good	Adequate
Guidelines ESCMID	CI	AI	BI (liposomal)
Cost	Very low	Medium	Very high (liposomal)

A: Strongly recommended; B: Moderately recommended; C: Marginally recommended; ESCMID: European Society of Clinical Microbiology and Infectious Diseases; I: Evidence from RCTs; II: Evidence from non randomized clinical trials; III: Evidence from expert opinions; RCTs: Randomized clinical trials.

The choice of an antifungal therapeutic strategy has a deep impact on the hospital epidemiology: Lortholary et al. reported that the extensive use of caspofungin decreased the isolation of C. albicans from 56 to 21% but increased C. glabrata from 18 to 35% and C. parapsilosis from 13 to 31% [25]. Similar epidemiological modifications were observed using fluconazole. The risk of infection with an isolate with decreased susceptibility to fluconazole or caspofungin is associated with the recent exposure to these drugs [26].

Although resistance to echinocandins is still generally uncommon, it has been reported to be on the rise. For instance, among 313 C. glabrata blood isolates, resistance to echinocandins increased from 4.9 to 12.3% in the 2001–2010 period [27].

A major concern for the extensive use of echinocandins is represented by cost.

Cost is definitely higher with respect to fluconazole but similar to the cost of true competitors. In particular, echinocandins cost a little bit more than voriconazole but less than liposomal amphotericin B.

Despite a higher cost, echinocandins (anidulafungin, micafungin) have demonstrated to reduce mortality and overall in-hospital costs compared to fluconazole, both in the setting of empirical and definite treatment of IC [28–30]. Notably, some authors suggest that caspofungin is more cost-effective than fluconazole in the empiric treatment of IC/C when fluconazole resistance is higher than 25% in hospital, reinforcing the importance of knowing the local epidemiology [31].

No difference between micafungin and caspofungin has been demonstrated in terms of cost–effectiveness in the treatment of candidemia and IC [32].

Caspofungin has been proved to be more cost-effective than liposomal amphotericin B in the empirical treatment of invasive fungal infections and the treatment of candidemia, not only for a lower cost of drug, but also for a lower incidence of renal failure [33,34]; similar data have been reported for micafungin [35].

De-escalation strategy (initial treatment with echinocandins, followed by fluconazole when possible) has been proved to reduce mortality and improve outcome of IC/C with a significant cost saving, compared to escalation strategy (initial treatment with fluconazole) [36].

However, a timely start of a broad-spectrum antifungal with fungicidal and anti-biofilm activity, switching to a cheaper alternative according to microbiology results and clinical status is considered to be the main determinant of cost–effectiveness, regardless of the antifungal agent [37].

In order to improve appropriateness of antifungal therapy, antifungal stewardship programs, as a sub-entity of antimicrobial stewardship, should be considered at least in larger institutions [38].

In conclusion, IC/C is a relevant cause of morbidity and mortality among inpatients. Early identification of patients at risk, knowledge of local epidemiology and prompt efforts to define etiologic diagnosis are pivotal to ensure appropriateness. Start with an echinocandin and switch to fluconazole when possible, seems to represent an useful strategy for the management of IC/C.

Financial & competing interests disclosure

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or pending, or royalties.

No writing assistance was utilized in the production of this manuscript.