Biofilm is a relevant risk factor for mortality in candidemia. Data about serum-based susceptibility testing against Candida biofilms are scant; therefore, the activity of fluconazole, amphotericin B, caspofungin and micafungin was determined against Candida albicans and C. parapsilosis biofilms with or without 50% human serum.
Ten clinical C. albicans and ten C. parapsilosis sensu stricto isolates derived from bloodstream were examined. Susceptibility testing was evaluated using XTT-based assays. The activity of antifungals was calculated using the following formula: fungal damage (%) = [1-(absorbance of experimental cells (492 nm)/absorbance of control wells (492 nm))]x100. The ratio of viable and dead cells was measured in fluorescent microscopic analysis using a LIVE/DEAD® BacLight™ viability kit.
Serum caused a remarkable adverse effect regarding biofilm structure for both species. Additionally, the ratio of non-viable cells increased for C. parapsilosis biofilms compared to C. albicans serum-exposed sessile populations. For fluconazole, significant differences were observed in susceptibility between serum-treated and serum-free biofilms in case of all C. albicans isolates at concentrations between 0.015 and 8 mg/L (p<0.01-0.05) (Fig1A). In case of amphotericin B, all isolates were more resistant in 50% serum at concentrations ranging from 0.5 to 1 mg/L as compared to susceptibility in serum-free RPMI-1640 (p<0.01-0.05) (Fig1C). In case of caspofungin, higher fungal damage was observed in RPMI-1640 compared to serum-containing medium at concentrations ranging from 0.015 mg/L to 1 mg/L for all isolates (p<0.05) (Fig1E). Micafungin showed high efficacy at concentrations from 0.015 mg/L (50.38%±13.09% to 83.0%±9.39%) in RPMI-1640 and from 2 mg/L (54.69%±10.68% to 86.25%±5.98%) in 50% human serum. At micafungin concentrations above 1 mg/L, there was no significant differences in fungal damage between serum-exposed and serum-free biofilms (p>0.05) (Fig1G). In case of C. parapsilosis, fluconazole did not result significant fungal damage at tested concentrations. However, enhanced activity was observed in 50% serum at concentrations from ≥0.03 mg/L (59.0%±7.24% to 99.63%±0.26%). Therefore, all isolates were more susceptible at concentrations ranging from 0.015 to 512 mg/L in 50% serum as compared to susceptibility in RPMI-1640 (p<0.01) (Fig1B). For amphotericin B, significant differences in metabolic activity were observed between RPMI-1640 and RPMI-1640 with 50% serum at concentrations between 0.015 and 1 mg/L (p<0.01) (Fig1D). The echinocandins showed high activity against C. parapsilosis biofilms in 50% serum, where significant metabolic activity decrease was observed at ≥0.015 mg/L (69.38%±11.53% to 99.38%±0.63% for caspofungin and 67.63%±11.37% to 99.63%±0.26% for micafungin, respectively) (p<0.01-0.05) (Fig1F,H). 
Though a better anti-biofilm activity is expected with echinocandins or amphotericin B, this is diminished by high protein binding. In contrast, the expectedly poorer activity of fluconazole on biofilms is enhanced by the inhibitory effect of serum on biofilm formation, which in turn does not decrease the effective fluconazole concentration markedly. Additionally, the significant biofilm mass decrease triggered by serum enhance the effect of fluconazole. The high antifungal activity of fluconazole in the presence of 50% serum both against C. albicans and C. parapsilosis biofilms support the usage of fluconazole as prophylaxis, which may reduce incidence of catheter-associated fungal infections especially in case of C. parapsilosis.

The effect of human serum on the efficacy of traditional antifungals against Candida albicans and Candida parapsilosis biofilms