In vitro activities of amphotericin B deoxycholate and liposomal amphotericin B against 604 clinical yeast isolates

We determined the in vitro antifungal activity of liposomal amphotericin B (L-AmB) against 604 clinical yeast isolates. Amphotericin B deoxycholate (D-AmB) was tested in parallel against all the isolates. Susceptibility testing was performed according to the Clinical and Laboratory Standards Institute (CLSI) M27-A3 method. Overall, L-AmB was highly active against the isolates (mean MIC, 0.42 µg ml−1; MIC90, 1 µg ml−1; 97.2 % of MICs were ≤1 µg ml−1) and comparable to D-AmB (mean MIC, 0.48 µg ml−1; MIC90, 1 µg ml−1; 97.3 % of MICs were ≤1 µg ml−1). The in vitro activity of D-AmB and L-AmB was correlated (R2 = 0.61; exp(b), 2.3; 95 % CI, 2.19–2.44, P<0.001). Candida albicans (mean MICs of D-AmB and L-AmB, 0.39 µg ml−1 and 0.31 µg ml−1, respectively) and Candida parapsilosis (mean MICs of D-AmB and L-AmB, 0.38 µg ml−1 and 0.35 µg ml−1, respectively) were the species most susceptible to the agents tested, while Candida krusei (currently named Issatchenkia orientalis) (mean MICs of D-AmB and L-AmB, 1.27 µg ml−1 and 1.13 µg ml−1, respectively) was the least susceptible. The excellent in vitro activity of L-AmB may have important implications for empirical treatment approaches and support its role in treatment of a wide range of invasive infections due to yeasts.


INTRODUCTION
Amphotericin B deoxycholate (D-AmB), a polyene macrolide, is the longest established antifungal agent and for many decades was considered the gold standard for the treatment of invasive fungal infections. It is active against many clinically relevant yeasts (i.e. Candida spp. and Cryptococcus neoformans) and moulds, including most of Aspergillus spp. and Mucorales (Adler- Moore & Proffitt, 2002;Lacerda & Oliveira, 2013;Moen et al., 2009). Acquired resistance to this agent is rare (Kanafani & Perfect, 2008).
The clinical use of D-AmB is impaired by its poor aqueous solubility and its toxicity, especially nephrotoxicity (nearly 50 % of patients), and by infusion-related reactions, such as fever and chills (Dupont, 2002). As a result of these limitations, in 2009 the Infectious Diseases Society of America (IDSA) guidelines (Pappas et al., 2009) introduced a significant change in the management of patients with invasive candidiasis: D-AmB, previously recommended as first-line therapy (Pappas et al., 2004), is now considered as an acceptable therapy only for invasive candidiasis in nonneutropenic patients intolerant or with limited access to other antifungal agents. To attenuate its adverse effects, lipid formulations of amphotericin B [liposomal amphotericin B (L-AmB), amphotericin B lipid complex and amphotericin B colloidal dispersion] were developed (Table 1). Several studies have indicated that the three lipid-based formulations are not therapeutically equivalent. L-AmB appears to be substantially less toxic than the other two formulations in terms of nephrotoxicity and incidence of infusion-related adverse events (Cifani et al., 2012;Enoch et al., 2006;Saliba & Dupont, 2008;Wade et al., 2013). Based on its enhanced safety and efficacy profile, the US Food and Drug Administration (www.fda.gov), the European Society of Clinical Microbiology and Infectious Diseases (ESCMID) (Ullmann et al., 2012) and the European Conference on Infections in Leukaemia (ECIL; Maertens et al., 2011) guidelines propose L-AmB for empiric antifungal therapy in febrile neutropenic patients. L-AmB is also recommended (AII recommendation) as therapy, with the same strength of recommendation as echinocandins, according to IDSA guidelines for neutropenic patients with candidaemia (Pappas et al., 2009). Moreover, whereas the clinical activity of L-AmB has been widely studied (Cifani et al., 2012;Enoch et al., 2006;Saliba & Dupont, 2008;Wade et al., 2013), there is not an extensive literature (Anaissie et al., 1991;Carrillo-Muñoz et al., 1999;Jessup et al., 2000;Lass-Flörl et al., 2008) concerning the in vitro susceptibility of this agent against yeasts.
The aim of this study was to investigate the in vitro susceptibility to L-AmB compared with D-AmB of clinically relevant yeasts obtained from critically ill and haematological patients with bloodstream infection. Comparisons between susceptibility testing results were undertaken in order to better understand the activity profile of L-AmB compared with D-AmB.
Prior to being tested, each isolate was subcultured on Sabouraud dextrose agar plates (bioMérieux) to ensure purity, viability and optimal growth characteristics.
Susceptibility testing. D-AmB (Sigma-Aldrich) and L-AmB (AmBisome; Gilead Sciences) were obtained as standard powders. Broth microdilution (BMD) testing was performed in accordance with the Clinical and Laboratory Standards Institute (CLSI) method M27-A3 (CLSI, 2008). Briefly, BMD panels containing serial twofold dilutions of each antifungal agent in RPMI 1640 medium (Sigma) buffered to pH 7.0 with MOPS (Sigma), frozen in 96-well plates at 280 uC for no more than 3 months, were thawed and inoculated with an organism suspension adjusted to attain a final inoculum concentration of 1.5610 3 ±1.0610 3 cells ml 21 . The final range of drug concentrations tested was 0.03-16 mg ml 21 . The minimum inhibitory concentration (MIC) values were visually determined, after 48 h of incubation at 35 uC, as the concentration that inhibited 100 % of fungal growth. The quality control isolates Candida krusei ATCC 6258 and Candida parapsilosis ATCC 22019 listed in CLSI (2008) were tested.
Analysis of results. CLSI has not determined breakpoints for amphotericin B. In order to perform a comparison in this study, the isolates inhibited by D-AmB or L-AmB at ¡1 mg ml 21 were considered susceptible, as detailed in a previous study (Diekema et al., 2009). 'Resistant' isolates were defined as isolates with MICs .1 mg ml 21 . MIC data are presented as the range, mean, MIC 50 (MIC causing inhibition of 50 % of isolates) and MIC 90 (MIC causing inhibition of 90 % of isolates) and for each species. MIC 50 and MIC 90 values were calculated for those species with 10 or more isolates.
To analyse the correlation between the two drugs, we built a simple regression model computing R 2 and exp(b) values with 95 % confidential interval (95 % CI). Moreover, to assess the differences in MICs among the species for each drug, the Kruskal-Wallis test was performed. The level of significance was set at a P-value ,0.05. All statistical analyses were carried out using STATA MP 11.2 for Mac Os X. Data were presented graphically using Excel. Table 2 summarizes the in vitro susceptibilities to the two different formulations of amphotericin B of 604 yeast isolates. Despite the differences in the structure, the spectra of activity of D-AmB and L-AmB are comparable. As shown by other authors (Anaissie et al., 1991;Carrillo-Muñoz et al., 1999;Jessup et al., 2000;Lass-Flörl et al., 2008), the two formulations demonstrate excellent potency and spectra: only 15 (2.5 %) and 16 (2.6 %) isolates had MIC levels that indicated resistance to D-AmB and L-AmB, respectively. D-AmB and L-AmB were statistically different D-AmB and L-AmB in vitro activities against yeasts

M. T. Montagna and others 1640
Journal of Medical Microbiology 63 with respect to yeast species (P50.001). Candida albicans (mean MICs of D-AmB and L-AmB, 0.39 and 0.31 mg ml 21 , respectively) and Candida parapsilosis (mean MICs of D-AmB and L-AmB, 0.38 and 0.35 mg ml 21 , respectively) appeared as the species most susceptible to the agents tested, in accordance with data from previous studies (Anaissie et al., 1991;Carrillo-Muñoz et al., 1999;Jessup et al., 2000;Lass-Flörl et al., 2008). Moreover, although Candida lusitaniae (Clavispora lusitaniae) is known to be intrinsically less susceptible to polyenes (Pappas et al., 2004), none of our strains exhibited MIC .1 mg ml 21 for either agent. Other authors (Lass-Flörl et al., 2008) have also reported a high susceptibility rate (98 %) by the EUCAST method [Subcommittee on Antifungal Susceptibility Testing (AFST) of the ESCMID European Committee for Antimicrobial Susceptibility Testing (EUCAST), 2008].
A decrease in the activity of D-AmB and L-AmB was noted among isolates of Candida glabrata (mean MICs of D-AmB and L-AmB, 0.75 and 0.72 mg ml 21 , respectively), as already reported by other investigators (Carrillo-Muñoz et al., 1999;González et al., 2008;Hull et al., 2012a;Pfaller et al., 2002). Recently published data (Hull et al., 2012b) demonstrate that missense mutation in ERG11 enables Candida glabrata to circumvent the inhibitory action of polyenes and azoles.
Regarding Candida krusei (Issatchenkia orientalis) susceptibility, MICs of D-AmB and L-AmB were found to be the highest (mean MICs of D-AmB and L-AmB, 1.27 and 1.13 mg ml 21 , respectively) as in other studies (Kiraz & Oz, 2011;Lass-Flörl et al., 2008;Pfaller et al., 2002;Ranque et al., 2012). However, the molecular mechanisms influencing the susceptibility of this species to polyenes are poorly understood.
The overall distribution of D-AmB and L-AmB MIC values is shown in Fig. 1 In conclusion, we have performed a head-to-head challenge of D-AmB and L-AmB against a large collection of clinical yeast isolates using the CLSI M27-A3 BMD method. The results of this study demonstrate high levels of inhibitory activity of L-AmB, though a reduced susceptibility was detected for Candida glabrata and Candida krusei (Issatchenkia orientalis). In addition, we also found a strong correlation between the in vitro antifungal activities of D-AmB and L-AmB. For this reason, we can assume that if resistance to one of the two agents emerges, it is reasonable to assume that the other agent will also show comparable results, although confirmation of this finding will need additional investigation. Based on our in vitro results and on comparative data from well-controlled trials and extensive clinical experience (Lacerda & Oliveira, 2013;Miceli & Chandrasekar, 2012;Moen et al., 2009) we conclude that, despite the availability of expanded-spectrum azoles and echinocandins, L-AmB remains a first-line drug choice for empirical therapy in patients with febrile neutropenia and it is also an option for the treatment of many patients with candidaemia. Constant surveillance is essential to monitor the activity of L-AmB against clinical yeast isolates in order to detect isolates with reduced susceptibility, thereby supporting the most appropriate choice of early antifungal treatment towards a better prognosis (Morrell et al., 2005;Taur et al., 2010).

ACKNOWLEDGEMENTS
This work was done at the Department of Biomedical Science and Human Oncology, Hygiene Section, University of Bari, Italy and was D-AmB and L-AmB in vitro activities against yeasts partially supported by a grant from Gilead Sciences, Milan, Italy. All authors declare that they have no conflicts of interest.