INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

A great deal of effort has gone into the development of the standardized method for antifungal susceptibility testing, and the reference techniques are now more reliable and reproducible (5, 10). This agreement is essential for identification of organisms unlikely to respond to certain antifungal treatments. The National Committee for Clinical Laboratory Standards (NCCLS) Subcommittee on Antifungal Susceptibility Testing has standardized testing methods for Candida spp. and Cryptococcus neoformans (M27-A;5) and proposed guidelines for filamentous fungi (M38-P;6). However, the use of the NCCLS methodology still has some limitations. In particular, a major limitation is the unreliability of detection of resistance to amphotericin B (AMB). RPMI medium yields a range of MICs that spans only 3 or 4 twofold serial dilutions. This short range precludes reliable discrimination between susceptible and resistant isolates (4, 7,10).

It has been suggested that the utilization of antibiotic medium 3 (AM3) instead of RPMI improves detection of resistance to AMB (9, 14). AM3 used as assay medium allows detection of both AMB-resistant Candida and Cryptococcus neoformans isolates (4, 8, 9). In addition, AM3 has shown enhanced ability to detect resistant isolates by both the dilution procedures of the M27-A reference method and the E-test agar diffusion method (8, 14, 15). However, the reproducibility of antifungal susceptibility testing (AST) using AM3 is still under study. The components of this medium are not completely defined, and substantial lot-to-lot variation has been detected (7, 9). Recent reports have pointed out that variability was found to be minimal with currently manufactured lots of AM3 (3). However, other studies have emphasized the potential for interlaboratory variability when AM3 is used (7). In addition, it is not clear if glucose supplementation of AM3 decreases or enhances the ability of AST using this medium to detect AMB resistance (4).

Variability of antimicrobial susceptibility results is a basic limitation for the reproducibility of MICs. Thus, some defined or semidefined synthetic media could be an alternative for detection of resistance to AMB (10). Iso-Sensitest is a semidefined medium for antimicrobial susceptibility testing in which undefined components are kept to a minimum (M. Cuenca-Estrella, J. L. Rodriguez-Tudela, T. M. Diaz-Guerra, and E. Mellado, Abstr. 40th Intersci. Conf. Antimicrob. Agents Chemother., abstr. J-922, p. 366, 2000). It was designed to overcome the lack of reproducibility with different peptones. It is made of an aminonitrogen base of acid-hydrolyzed casein and special peptones supplemented with defined growth factors including metal ions. Iso-Sensitest medium allows the growth of the great majority of microorganisms without further supplementation (2, 13).

The aim of this study was to assess the abilities of Iso-Sensitest to distinguish amphotericin B-susceptible and -resistant isolates. Results obtained for two different lots of Iso-Sensitest broth as assay medium for AST were compared with those obtained by the NCCLS reference methodology. Results were also compared with both those obtained using AM3 and those employing RPMI-2% glucose as assay medium for AST. The influences of inoculum, glucose supplementation, reading method, and incubation period on MICs of AMB were also analyzed.

(This work was presented in part at the 40th Interscience Conference on Antimicrobial Agents and Chemotherapy, Toronto, Canada, 2000 [Cuenca-Estrella et al.].)

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Organisms. A collection of previously described putatively AMB-susceptible and -resistant isolates was used (3, 9, 15). Susceptible isolates consisted of four Candida albicans strains (CNM-CL-3414 [Centro Nacional de Microbiologia yeast culture collection], CNM-CL-3416, CNM-CL-3419, and CNM-CL-3420), three Candida parapsilosis strains (ATCC 200954, CNM-CL-3413, and CNM-CL-3417), and one Candida lusitaniae strain (CNM-CL-3418); resistant isolates consisted of four C. lusitaniae strains (ATCC 200950, ATCC 200951, ATCC 200952, and ATCC 200953), one C. albicans strain (ATCC 200955), and one Candida tropicalis strain (ATCC 200956). Putatively susceptible isolates either were obtained from the bloodstream from patients who responded to treatment with AMB or had been proven to be susceptible in animal models. Putatively resistant organisms had been proven to be resistant in animal models or showed elevated (>4 µg/ml) MICs of AMB in numerous susceptibility tests.

AST. Table 1 lists the methods evaluated. Two different lots of each medium were employed. Media were sterilized by filtration and prepared as a double-strength solution. RPMI, AM3, and Iso-Sensitest were prepared according to the manufacturers' instructions, and the pH was adjusted to 7.0 with NaOH or HCl when necessary. Drug stock preparation, storage conditions, dilution techniques, trays, and inoculum preparations were performed by the NCCLS reference microdilution method (5). AMB was obtained as standard powder from Sigma Aldrich Química, Madrid, Spain.

Endpoint determination. For visual MIC determination, endpoints were defined as the lowest concentration of AMB that completely inhibited the growth of the strain. For spectrophotometric endpoint determination, the MICs were obtained by measuring the absorbance at 540 nm with a Labsystems IEMS Reader MF (Labsystems, Madrid, Spain). The spectrophotometric MIC of AMB was defined as the lowest drug concentration leading to a 90% or greater growth reduction compared with the drug-free control (1). Spectrophotometric reading allowed the determination of the growth index for each medium-inoculum-incubation time combination. The growth index was measured relative to the optical density at 540 nm of the drug-free well serving as the growth control.

Statistical analysis. The optical density in the drug-free well must be >0.2 to calculate the spectrophotometric MICs. The mean absorbance of eight sterility control wells was subtracted from the absorbance measured for each well, and then spectrophotometric MICs were calculated. The significance of the differences between methods was determined by the Student t test (unpaired, unequal variance) or by the Mann-Whitney U test. When the effect of one variable was studied, the others were fixed as constants. The correlation between methods was determined by Pearson's coefficient expressed over a maximum value of 1. MICs were transformed to log₂ data. A P value of <0.05 was considered significant.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

RPMI. Tables 2 and 3 summarize the AST results obtained by visual reading and spectrophotometric reading, both after 24 and 48 h of incubation. The tables list in vitro results for each medium-inoculum combination. It can be noted that AST methods employing RPMI (with or without glucose) as assay medium failed to detect resistance to AMB. The MIC ranges for the putatively susceptible and resistant isolates showed frequent overlaps. No significant differences were found between visual and spectrophotometric MICs (P > 0.05). Incubation period and inoculum size did not have a significant influence on MICs either.

AM3. In general, the results for AM3 with or without glucose were highly correlated with those obtained with RPMI (Pearson's r coefficients > 0.85). However, the ranges of observed MICs for putatively susceptible and resistant isolates did not overlap when an inoculum size of 10³ CFU/ml was employed, permitting discrimination between groups. A reliable discrimination was not obtained with AM3 and inocula of 10⁵ CFU/ml. A significant inoculum effect was observed (P < 0.05). Ranges for MICs obtained using an inoculum size of 10⁵ CFU/ml were more elevated than those obtained with the lower inoculum. For inocula of 10³ CFU/ml, Tables 2 and 3 show that MICs determined spectrophotometrically and those determined visually, after 24 and 48 h of incubation, were in similar ranges. Glucose supplementation did not significantly influence MICs (P > 0.05).

Iso-Sensitest. Methods using Iso-Sensitest as assay medium yielded AMB MIC ranges broader than both those obtained with RPMI and those obtained with AM3. In particular, putatively susceptible isolates demonstrated lower MICs after 24 h of incubation and by spectrophotometric reading than those obtained with other assay media. The results achieved with Iso-Sensitest alone and Iso-Sensitest with glucose showed good correlation indexes compared with both those obtained with RPMI and those obtained with AM3 (r > 0.78 and r > 0.83, respectively). Reading method had no significant influence on MIC determination. However, glucose supplementation, incubation time, and inoculum size significantly influenced MICs of AMB (P < 0.05). In short, an incubation time of 24 h and an inoculum of 10³ CFU/ml yielded significantly lower ranges for MICs than those encountered after 48 h of incubation with an inoculum of 10⁵ CFU/ml. The MICs for susceptible isolates fell when AST was performed with Iso-Sensitest supplemented with glucose, and these methods showed enhanced abilities to discriminate between putatively susceptible and resistant isolates (Table 4).

Growth levels for different medium-inoculum-incubation time combinations. Table 5 lists growth indexes for the 14 putatively AMB-susceptible and -resistant isolates and the two ATCC strains. For each assay medium analyzed, glucose supplementation and an inoculum size of 10⁵ CFU/ml had additive effect on the growth index (P < 0.05). The best indexes of growth were achieved with Iso-Sensitest supplemented with glucose. The AM3 medium supplemented with glucose also yielded luxuriant growth. The lowest growth indexes were obtained with the NCCLS reference method (RPMI with an inoculum of 10³ CFU/ml).

Reproducibility. Table 6 shows the degree of reproducibility for each method. The table displays results of nine consecutive MIC determinations for the two ATCC strains. The ICCs are a reverse measurement of the lot-to-lot variability of the medium and are expressed over a maximum value of 1. The reproducibility of AMB MICs was high for each method (ICC > 0.723). In general, lower reproducibility was observed with visual reading than with spectrophotometry. The ICCs of results for AM3 alone and AM3 with glucose were lower than both those obtained with RPMI alone or with glucose and those obtained with Iso-Sensitest alone or with glucose (P < 0.05). The 95% confidence intervals for AM3 were the broadest, indicating a lower reliability. The highest ICCs were obtained for Iso-Sensitest (ICC > 0.949).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Despite its limitations the NCCLS reference method for AST of yeasts is a new milestone in the evolution of medical mycology. The good interlaboratory reproducibility of the NCCLS reference methods has led to the establishment of interpretative breakpoints for fluconazole, itraconazole, and flucytosine (10). Tentative breakpoints for AMB have not been proposed, however. Evidence of correlation between the clinical outcome and AMB susceptibility results is limited due to the lack of a reliable means of detecting resistance to this antifungal agent employing the reference method (5). Some reports point out that results obtained by E-test or minimal fungicidal concentration determination are better predictors of clinical failure of AMB therapy than dilution reference procedures (4, 7, 8, 15). However, these methods are not in NCCLS document M27-A, and a more reliable detection of resistance to AMB is clearly needed.

Several works suggest that with AM3 as assay medium a more reliable detection of isolates showing AMB resistance is achieved (9). Unfortunately, this medium is not a standardized medium containing yeast extract, beef extract, and peptone, and lot-to-lot variation is thus possible (3). Since undefined media can give rise to widely varying results, the reproducibility of AST using AM3 is still under study. Some reports have indicated that AM3 supplemented with glucose enhanced the ability of the medium to detect resistance to AMB in Candida isolates (4, 15). Others found that glucose supplementation decreases the ability of AM3 to detect AMB resistance. In addition, it is not clear if results obtained after 24 h of incubation are more accurate than those obtained after 48 h (8).

A completely synthetic medium is usually recommended for susceptibility testing. Iso-Sensitest is a semidefined synthetic medium for antimicrobial susceptibility testing (Cuenca-Estrella et al., 40th ICAAC). The aim of this work was to test the abilities of Iso-Sensitest to detect AMB resistance in isolates for which the clinical correlation was known. The results obtained with Iso-Sensitest were compared with both those obtained with AM3 and those obtained with RPMI. In addition, the combined effect of glucose supplementation, reading method, and inoculum size was also evaluated.

As have other reports, we have confirmed that the NCCLS reference procedure has a reduced ability to detect AMB resistance (3, 9, 14). The addition of glucose, a higher inoculum size, or the two modifications together do not help in identifying AMB-resistant Candida isolates. Neither spectrophotometric reading nor visual reading permits detection of AMB resistance. Glucose supplementation and a larger inoculum size show an additive effect on the growth index. This method presents the advantage of reducing incubation time, but AMB resistance is not reliably identified.

In this report, AM3 medium shows an enhanced ability to detect putatively AMB-resistant organisms. Supplementation with glucose and reading method do not have a significant influence on MICs obtained with AM3. However, significant differences are observed in the growth index. Denser growth indexes are obtained when the medium is supplemented with glucose than without glucose. In particular, the high growth index (0.73 + 0.23 U) obtained with AM3-2% glucose and 10³ CFU/ml after 24 h of incubation indicates that the incubation period needed is shorter than that needed without glucose. In contrast, an inoculum size of 10⁵ CFU/ml does not allow detection of resistance. A significant inoculum effect is observed when an inoculum of 10⁵ CFU/ml is used. Thus, the results are falsely elevated, even the ranges for MICs of putatively susceptible isolates, and a reliable discrimination is not obtained when AM3 and an inoculum of 10⁵ CFU/ml are employed. The ICCs for AM3 are elevated (>0.723), although less than those achieved with both RPMI and Iso-Sensitest. In general, variability of results with AM3 is minimally significant, and a reliable detection of resistance to AMB is observed. As previous studies suggest, isolates for which MICs are 0.25 µg/ml with AM3 medium should be classified as resistant (3, 4).

Finally, the reproducibility of results with Iso-Sensitest is very elevated, as expected since this medium is synthetic. Each AST procedure performed with Iso-Sensitest yielded MIC ranges broader than those obtained with AM3 and RPMI. However, a reliable detection of resistance to AMB was seen with ASTs employing Iso-Sensitest-2% glucose only. In particular, the discrimination is better after 24 h of incubation and by spectrophotometric reading. In addition, the growth indexes with glucose are higher than without glucose, permitting a reduction in the incubation time. Further, the elevated correlation observed between results obtained with Iso-Sensitest and both RPMI and AM3 suggests that the susceptibility testing done with Iso-Sensitest does not produce an overall shift in MICs but rather reliably discriminates between AMB-susceptible and -resistant organisms.

In summary, the data presented demonstrate that the glucose supplementation, a higher inoculum size, the spectrophotometric endpoint determination, and reading after an incubation period of 24 h do not enhance the abilities of AST methods using RPMI as assay medium to detect AMB resistance. Secondly, susceptibility testing employing AM3 with an inoculum of 10³ CFU/ml distinguishes between resistant and susceptible isolates. The reading method, incubation time, and addition of glucose do not have significant effect on MICs. However, an inoculum size of 10⁵ CFU/ml shows a significant inoculum effect, and a reliable discrimination is not achieved. Finally, Iso-Sensitest as assay medium allows detection of AMB resistance. The abilities of AST using this medium are similar to those of methods employing AM3, and reproducibility is higher. Because of the small number of strains tested these findings require further investigation.