Panel of 44 Aspergillus isolates from clinical samples like sputum, nasal mass, endotracheal aspirate, blood cultures, deep seated pus, brain tissue, pleural and peritoneal fluid,were tested over a period of two years (2011-2013) in the Department of Microbiology, King George’s Medical University, Lucknow, U.P, India. These isolates comprised of 34 A.fumigatus, 8 A. flavus, 2 A.terreus. Two reference strains, Candida krusei ATCC 6258 and A. flavus ATCC MYA 3631 were included in all the tests as references for quality controls and were tested by all the methods in each series of assays.

Inoculum Preparation: Inoculum was prepared according to M51-A and M38-A2 CLSI guidelines. In brief to induce conidium the isolates were grown on potato dextrose agar (PDA) slants at 35^oC for 2–7 days before testing. The suspensions of conidial inocula were prepared from 2-7 days old cultures grown on PDA slants. The turbidity of the cell suspension was adjusted by spectrophotometry to an optical density of 0.09 to 0.13 for Aspergillus spp.

Broth Micro-dilution (BMD): Voriconazole (Cayman, USA), Amphotericin B (BioVision, USA) and Caspofungin (Cayman, USA) were obtained as powders. Stock inoculum and working suspensions for all the drugs were prepared and BMD was done according to the method described in CLSI M38-A2 document [4]. All BMD plates were incubated at 35^oC for 48 h. MIC for AMB and VCZ and MEC for CAS were interpreted in accordance with ECV values suggested by CLSIM38-A2. Results were read at 48 h for VCZ and AMB and at 24 h for CAS. In brief, isolate was considered as wild type (WT) when MIC was ≤ 1μg/ml for VCZ (≤ 2 μg/ml for A.flavus), ≤ 2 μg/ml for AMB (for A. fumigatus, A. flavus, A.niger and ≤ 4 μg/ml for A. terreus) and MEC was ≤ 0.06 μg/ml for CAS [10–13].

Disk diffusion (DD): Antifungal susceptibility by disk diffusion was done on non-supplemented MHA according to the method described in CLSI M 51-A document [7]. In brief entire surface of MHA was inoculated with non-toxic cotton swab with the undiluted mold stock inoculum suspension. Disks of AMB (10μg, prepared Inhouse); VCZ (1 μg, Himedia); CAS (5μg, In house) were placed onto the surface of each of the inoculated MHA plate. The plates were incubated at 35^oC and were read after 24h and 48h. Slight trailing around the zone edge or hyphal element extending into the inhibition zone was ignored when testing triazoles, but not for AMB. Microcolonies or trailing growth within a well-defined zone of inhibition was ignored when testing CAS.Zone diameter were interpreted in accordance with CLSI M51-A. In brief for AMB isolates were considered as WT at zone size ≥ 15 mm. For VCZ and CAS, isolates were considered as WT at zone size ≥ 17 mm [7].

E-test: The E-test strips (Biomeriux, USA) of AMB, VCZ and CAS contained concentration gradients of 32–0.002 μg/ml of the respective drug. The entire surface of non-supplemented MHA was inoculated by the same method as used for DD. E strips were placed on MHA and the plates were incubated at 35^oC and read after 24h and 48h. The MIC was determined from the inhibition ellipse that intersected the scale on the strip. The MIC was read at the point of significant inhibition in fungal growth. Micro colonies or trailing growth within a well-defined zone of inhibition ellipse was ignored when testing CAS [13].

Data Analysis: For comparative evaluation of the disk diffusion, E-test and broth micro-dilution methods, the geometric mean (GM) and range of the MICs and MECs and the range of the inhibition zone diameters were calculated for each genus-species combination. The percentage of agreement between all the methods was calculated for each genus-species and antifungal drug.

All clinical isolates grew well in the RPMI media used for BMD and non-supplemented MHA media used for DD and E-test, giving definitive endpoints. The MIC and zone diameter ranges and geometric means obtained by BMD, E-test and DD are summarized in [Table/Fig-1].

Antifungal Susceptibility

Based on ECVs, Aspergillus isolates were classified into WT and non-WT for all three antifungals tested [Table/Fig-2].

Eight WT isolates of A.fumigatus by BMD were categorized as non-WT by DD. Also, 6 non-WT isolates by BMD were categorized as WT by DD. Among these 6, 4 were A. fumigates and 2 were A. flavus isolates. In comparison to BMD lower MIC values by E-test were seen in 4/44 (9.1%) Aspergillus isolates (2 A. flavus and 2 A. fumigatus).

Four isolates categorized as non-WT (MIC’s > 1 μg/ml) by BMD were categorized as WT by DD and E-test. Among these 2 were A. fumigatus, 2 were A. flavus. MIC values by E-test for these 4 isolates were < 1 μg/ml.

The percentage of agreement of E test and DD with the gold standard BMD in different species for the three antifungals is summarized in [Table/Fig-3]. Percent agreement of susceptibility testing by DD and E-test ranged from 75% to 100 %, being lowest for A. flavus. When testing for AMB and VCZ, the overall agreement of E test with BMD was better than that of DD [Table/Fig-4]. 100% agreement was seen between all the three methods for CAS.

For 20 isolates BMD MIC values obtained were different at 24h vs 48h. Of these, 14 isolates whose MIC were within ECV, the interpretation at 24 h gave still lower MIC’s. For 6 isolates, which were non-WT at 48h, the interpretation at 24h gave MIC within ECV (WT).There was no difference in the results of either DD or E-test at 24h or 48 h.

When the BMD results were interpreted at 24h, the agreement of both E test and DD decreased for AMB, increased for VCZ while no change was seen for CAS [Table/Fig-4,5].

CLSI in year 2010 published disk diffusion method of antifungal susceptibility of non-dermatophyte filamentous fungi (CLSI M51-A). This method promises to be convenient for use in routine diagnostic laboratories. Though the fungal infections are on rise in India, data on antifungal susceptibility especially for filamentous fungi are lacking.

In this study high MIC value of ≥ 4 μg/ml for amphotericin B was found in 31.8% isolates of Aspergillus spp. High MIC values (≥ 2 μg/ml) had also been reported by Shivprakash et al., in 3 isolates of A.flavus [14]. The trend in increase in non-WT (probable resistant) strains of Aspergillus in India is now on rise and this trend could be because of rampant use of amphotericin B in the patients suspected of fungal infections. There is a possibility of increase in MIC values during therapy, as has been noted by Espinell-Ingroff et al., for a single A. fumigatus isolate [15].

The mechanism of resistance in A. flavus has been reported to be due to efflux of drug or increased transcription of AflMDR1 gene, alteration of cell wall composition [16]. The mechanism of the intrinsic resistance in A.terreus to amphotericin B is not clear and it has been postulated that much less ergosterol content in the cell membrane of A. terreus partially accounts for the poor activity of amphotericin B against this fungus [17]. One of the study also reported that A. terreus produces significantly more catalase than A.fumigatus which may play an important role in the amphotericin B resistance [18].

We also found voriconazole MIC of ≥ 2 μg/ml in 9.1% Aspergillus isolates. Voriconazole MIC of ≥ 4 μg/ml had also been reported by Shivprakash et al., in 4.9% of A. flavus [14]. Espinell-Ingroff et al., also reported voriconazole MIC exceeding 4 μg/ml in 5.8% of the isolates, including one A. fumigatus, 7.7% F. solani and all zygomycetes [15]. A study from Spain found voriconazole MIC of > 4 μg/ml in 10 A. fumigatus clinical isolates collected from patients with hematological malignancies who had long-term exposure to either voriconazole or Itraconazole [19]. Resistance of A. fumigatus to the azoles has been reported to vary from a high of 52% and 38% with itraconazole and ravuconazole, respectively, to a low of 11% with voriconazole [20]. The studies on the resistant mechanisms have shown that amino acid residues substitution derived from mutations in the azole-target-enzyme gene cyp51A, overexpression of this gene and drug efflux genes, and up regulation of homeostatic stress-response pathways contribute to azole-resistance in A. fumigates [21]. Liu W et al., identified that T788G missense mutation in cyp51C gene was responsible for voriconazole resistance in A. flavus [22].

None of the Aspergillus isolate was found to have MEC values above ECV’s for caspofungin. These findings are similar to the findings of Shiv prakash et al., who also could not find high MEC values in any of the A. flavus [14]. Resistance to caspofungin in Aspergillus had not been reported from India. This could be because caspofungin is still the least used antifungal in India. Arendrup et al., had reported a clinical isolate of A. fumigatus to be resistant to caspofungin, confirmed by E-test with an MIC >32 μg/ml. This patient also failed to respond to caspofungin [23]. Laboratory-selected strains with varying degree of caspofungin resistance have been described. Some of these laboratory-manipulated strains have been found to have mutations in the ECM33 gene (afuEcm33), a gene which encodes cell wall protein important for fungal cell wall organization [24].

E-test has been evaluated by various authors and most of the authors have used RPMI 1640 agar with 2% glucose and 0.165 M morpholine propane sulfonic acid (MOPS) buffer. We evaluated E-test on non-supplemented MHA as is being used for DD method of CLSI. Espinell Ingroff et al., used RPMI 1640 agar and found agreement between E-test and BMD ranging from 64.5% to 96.3% for voriconazole, posaconazole, itraconazole and amphotericin B [13]. Since agreement in our study for E-test ranged from 75 to 100% for Aspergillus spp. and also since there was no difference in MIC or MEC values at 24h and 48 h, therefore we suggest E-test to be done on non-supplemented MHA in place of RPMI 1640 agar with 2% glucose and MOPS, which is also the common media available in most of the laboratories.

CLSI M38-A2 guidelines recommend BMD interpretation at 48 hours of incubation for amphotericin B and voriconazole and at 24h of incubation for caspofungin. In this study comparison between methods for assessing the susceptibilities of Aspergillus spp. against amphotericin B, voriconazole and caspofungin showed that overall agreement after 48 h of incubation was highest for voriconazole (93.8% by both E-test and DD) and caspofungin (100%by both E-test and DD). Agreement was lowest for amphotericin B (93.8% by E-test and 87.5% by DD) at 48 h of incubation. But if the BMD interpretation is done at 24 h the agreement between all the methods increases specially for voriconazole (95%by both E-test and DD) but decreases considerably for amphotericin B [Table/Fig-5].We therefore suggest BMD interpretation of voriconazolealso at 24h in place of 48h as suggested by Espinell-Ingroff et al., [10].

Among Aspergillus spp. least agreement was found in susceptibility testing of A. flavus for amphotericin B and voriconazole by both E-test and DD [Table/Fig-3]. DD method of susceptibility testing of A. fumigatus for amphotericin B also showed lower agreement when compared to BMD [Table/Fig-3]. All MEC values by BMD for caspofungin were same, both at 24h and 48h. This finding is similar to the findings of Espinell-Ingroff et al., [11].

Global trend of increasing antifungal resistance in Aspergillus spp. is also being seen in India. CLSI method of disk diffusion must be interpreted with caution in case of amphotericin B susceptibility testing. MIC or MEC values do not change significantly for VCZ when BMD is interpreted at 24h in place of 48h as suggested by CLSI, infact BMD MIC values at 24h have better correlation with DD and E-test. We also propose E-test to be done on non-supplemented MHA media in place of RPMI 1640 agar. To increase the reproducibility of data large sample size and multi-centric studies are required.