Nitroglycerin-Citrate-Ethanol Catheter Lock Solution Is Highly Effective for In Vitro Eradication of Candida auris Biofilm

Candida auris poses emerging risks for causing severe central line-associated bloodstream infections. We tested in vitro the ability of antifungal lock solutions to rapidly eradicate C. auris biofilms.

ABSTRACT

Candida auris poses emerging risks for causing severe central line-associated bloodstream infections. We tested in vitro the ability of antifungal lock solutions to rapidly eradicate C. auris biofilms. Liposomal amphotericin B, amphotericin B deoxycholate, fluconazole, voriconazole, micafungin, caspofungin, and anidulafungin failed to completely eradicate all 10 tested C. auris biofilms. Conversely, nitroglycerin-citrate-ethanol (NiCE) catheter lock solution completely eradicated all replicates for all of C. auris biofilms tested.

TEXT

The importance of central lines as sources of bloodstream candidiasis has steadily risen over the last 3 decades (1–4). Measures to prevent candida central line-associated bloodstream infections (CLABSIs) have been implemented and include the use of prophylactic azoles, echinocandins, and amphotericin for high-risk patients (5). The reported prevalence of Candida auris infection has increased since 2009 (6, 7), and it has a tendency to be resistant to antifungals (8).

The luminal surfaces of central lines are vulnerable to C. auris colonization in high-risk patients and would benefit from effective prophylaxis (9, 10). We assessed the potential of several antifungal lock solutions suitable for intravascular flushing for disinfecting C. auris biofilms within time frames practical for use in central lines in cancer patients and compared those with nitroglycerin-citrate-ethanol (NiCE) lock solution.

Ten strains of C. auris (AR0381 to AR0390) were obtained from the Centers for Disease Control and Prevention (CDC) Antibiotic Resistance Isolate Bank. All isolates have been well characterized by the CDC, including susceptibilities and sequencing (11).

All antifungal drug lock solutions were tested at maximum infusion doses suitable for flushing: caspofungin 0.5 mg/ml (Cancidas Worldwide, Merck & Co., Inc., NJ) (12), micafungin 0.5 mg/ml (Mycamine, Astellas, Tokyo, Japan) (13), anidulafungin 0.5 mg/ml (Eraxis, Pfizer, New York, NY) (14), voriconazole 0.5 mg/ml (VFend IV, Pfizer) (15), fluconazole 2 mg/ml (Diflucan, Pfizer) (16), liposomal amphotericin B 1 mg/ml (Ambisome, Gilead, San Dimas, CA) (17), and amphotericin B 0.1 mg/ml (Fungizone, X-GEN Pharmaceuticals, Big Flats, NY) (18). NiCE was prepared from individual components, 0.003% nitroglycerin (Baxter Healthcare Corporation, Deerfield, IL), 4% disodium citrate (Sigma-Aldrich, St. Louis, MO), and 22% ethanol (Sigma-Aldrich).

Assessments of lock solutions were conducted using a well-established biofilm eradication model (19, 20). Briefly, six silicone discs colonized with 24 h biofilm were exposed to each lock for 2 h, sonicated to disrupt biofilm, and quantitatively cultured to enumerate the remaining viable organisms. The Kruskal-Wallis test was used to assess overall differences between strains and between locks. The Mann-Whitney U test was used for pairwise comparisons of NiCE with antifungal locks. All tests were two sided, with an α level of 0.05.

Although there were overall significant differences among C. auris isolates (P = 0.001), most strains had consistent (within 1 log₁₀) biofilm growth within each strain for the six replicates tested (Fig. 1). However, AR0382 and AR0385 demonstrated variability in biofilm densities that ranged from 10³ to 10⁵ CFU/disc. Quantitative results for efficacy of lock solutions against C. auris (AR0381 to AR0390) are presented in Fig. 2. Micafungin, anidulafungin, fluconazole, and voriconazole results were not significantly different (P > 0.05) from those of control for all strains. Liposomal amphotericin B showed complete kill for only 1 of 10 strains (AR0384) of C. auris. AR0387 had reduced susceptibility in biofilm and was not fully eradicated in four out of six replicates. Nonliposomal amphotericin B, in contrast, accomplished complete kill against 3 of the 10 strains (AR0381, AR0384, and AR0385) and partial kill for the remaining 7 strains (P < 0.05). NiCE and caspofungin were significantly more effective in eradicating C. auris biofilms than the positive control (P = 0.002 and 0.008, respectively). Caspofungin failed to completely eradicate all replicates from strains AR0386 to AR0389, whereas NiCE completely eradicated all replicates from the 10 strains.

FIG 1.

Variability of biofilm formation between strains of C. auris. C. auris biofilm demonstrated significantly different (P = 0.001) biofilm densities between the 10 strains tested. Within each strain, most replicates had consistent densities (within 1 log₁₀); however, AR0382 and AR0385 showed >2 log differences within the 6 replicates.

FIG 2.

Effectiveness of antifungal lock solutions after 2 h exposure against C. auris (AR0381 to AR0390) biofilm. Data are presented as median ± range for each strain of C. auris AR0381 to AR0383 (top), AR0384 to AR0386 (middle), and AR0387 to AR0390 (bottom). Only NiCE (0.003% nitroglycerin [GTN], 4% citrate, 22% ethanol [EtOH]) completely eradicated all C. auris strains significantly (P = 0.002) versus control. Caspofungin (0.5 mg/ml) was also significant (P = 0.008) versus control for all strains but failed to eradicate all replicates of all strains. Furthermore, amphotericin B (0.1 mg/ml) demonstrated significant (P < 0.05) effectiveness compared with control but failed to eradicate all replicates against AR0386 (middle) and AR0387 to AR0390 (bottom). None of the other solutions tested showed a significant difference compared with control (P > 0.05). * indicates significance compared to positive control.

In this study, a series of prophylactic antifungal catheter flush solutions were tested against several C. auris biofilm strains. In our hospital, locked catheter lumens are typically flushed intravascularly, and for high-risk patients, lumens can reliably be removed from treatment use for only a couple of hours. Testing was performed on 24 h C. auris biofilms followed by 2 h exposure to the lock solutions to represent a worst-case scenario for CLABSI prevention. For NiCE lock solution, the concentration used in our study was well tolerated in a phase 2 clinical trial where patients with hematologic malignancies underwent repeated catheter flushes (21). Furthermore, the same concentration was found to be highly effective in eradicating a wide spectrum of resistant bacteria and fungi (Candida) in biofilms (20, 22).

Fluconazole and voriconazole locks were not effective against C. auris biofilms (Fig. 2). This is consistent with planktonic C. auris resistance against fluconazole (>32 μg/ml). In contrast, voriconazole was effective against planktonic C. auris isolates (MIC, 0.5 to 4 μg/ml) (11). In biofilms, azole resistance may be reflective of efflux pump upregulation (23, 24), and multiple genes encoding efflux pumps are present across the C. auris genome (25). In addition, all planktonic C. auris strains tested were susceptible to micafungin (MIC, 0.125 to 0.5 μg/ml) and anidulafungin (MIC, 0.125 to 2 μg/ml) (11), yet they failed to eradicate C. auris biofilm (Fig. 2). As glucan synthesis inhibitors, echinocandins are known to lose efficacy due to mutations in Candida glucan synthase structure (26). C. auris strains have been reported to carry FKS mutations, and this genotype has correlated with echinocandin nonsusceptibility (27).

Liposomal amphotericin B was effective against some C. auris strains (AR0382, AR0383, and AR0384) but had only limited effectiveness against the others. This may reflect an inability of the phospholipid shell of encapsulated amphotericin to penetrate C. auris biofilms. The extracellular matrix of C. auris, although not fully elucidated, has been reported to be between Candida albicans, which secretes significant extracellular matrix and possesses significant hyphal structure, and Candida glabrata, which has little extracellular matrix consisting mainly of cells (28). Based on the targeting of liposomal amphotericin B to the fungal membrane, its reduced effectiveness compared with non-liposome-encapsulated amphotericin B may reflect an inability to rapidly reach yeast cell membranes in C. auris biofilms (29).

The tendency of C. auris to form biofilms renders central lines at high risk for colonization and for candidemia when C. auris is present in clinical settings. Prophylactic use of systemic antifungal drugs presents the potential for inducing antifungal resistance. NiCE appears to be an attractive option for further clinical study in preventing C. auris CLABSIs without presenting this risk. The effectiveness of NiCE for prevention of C. auris CLABSIs remains to be verified in clinical trials.