Previous Article | Next Article ![]()
Journal of Clinical Microbiology, August 2004, p. 3589-3593, Vol. 42, No. 8
0095-1137/04/$08.00+0 DOI: 10.1128/JCM.42.8.3589-3593.2004
Copyright © 2004, American Society for Microbiology. All Rights Reserved.
Mycology Reference Laboratory (Hellenic Centre for Diseases Control), Department of Microbiology, Medical School, University of Athens, Athens, Greece
Received 30 January 2004/ Returned for modification 28 March 2004/ Accepted 20 April 2004
| ABSTRACT |
|---|
|
|
|---|
16 µg/ml) against M. furfur, M. restricta, M. globosa, and M. slooffiae strains, which were AM3 confirmed. Agreement of the two methods was 84 to 97%, and intraclass correlation coefficients were statistically significant (P < 0.001). Because of higher amphotericin B MICs provided by Etest for strains also displaying high BMD MICs (
1 µg/ml), agreement was poorer. The proposed media are used for the first time and can support optimum growth of eight Malassezia species for recording concordant BMD and Etest MICs. | INTRODUCTION |
|---|
|
|
|---|
As the incidence of yeasts in deep-seated infections continues to increase in proportion to the growing numbers of immunocompromised, cancer, and postoperative patients, standardized MIC testing for more yeast genera can become pertinent to clinical practice. Regarding bloodstream Malassezia infections, poor response to amphotericin B has led to discontinuation of parenteral therapy as an additional therapeutic measure (18) or modification of therapy in order to increase the amphotericin B concentration in the catheter lumen, aiming at eradicating Malasssezia yeasts (1). However, the observed poor clinical response (10, 18, 19, 21) has not been supported by in vitro data, as susceptibility of Malassezia furfur (sensu lato) to amphotericin B has been recorded (9). Apart from host and yeast factors responsible for the in vivo-in vitro difference in amphotericin B efficiency, the effect of medium composition on yeast growth and drug availability may have accounted for the observed discrepancies between laboratory and clinical findings.
The aims of our study were to adapt the NCCLS M27-A2 microdilution reference method (14) for testing the susceptibilities of eight Malassezia species (5, 22) to amphotericin B, itraconazole, fluconazole, voriconazole, ketoconazole, and terbinafine to the new triazole posaconazole and to formulate a suitable solid medium to achieve concordant microdilution and Etest results.
| MATERIALS AND METHODS |
|---|
|
|
|---|
|
(i) BMD method. Susceptibility testing using RPMI 1640 (Sigma) as a diluent (14) supplemented, per liter, with 20 g of glucose, 4 g of ox bile (Oxoid), 1 ml of glycerol (Sigma), 0.5 g of glycerol monostearate (Sigma), and 0.4 ml of Tween 20 (Sigma). Itraconazole, ketoconazole, fluconazole, voriconazole, posaconazole, and amphotericin B suspensions and the final concentrations of antifungal drugs were prepared according to the NCCLS M27-A2 guidelines (14).
All stock inoculum suspensions were prepared in 5 ml of 50% (vol/vol) sterile saline-Tween 20 (Sigma) containing sterile glass beads 1.1 to 1.2 mm in diameter (Sherwood, St. Louis, Mo.). After 20 s of vortexing to disperse Malassezia clumps, the inoculum was standardized spectrophotometrically at 530 nm and adjusted with the modified RPMI medium to a final inoculum size of 2.0 x 103 to 3.5 x 103 CFU/ml, as validated by quality control (QC) quantitative plate counts of CFU in LN agar. Quantitative plate counts were not attainable with all Malassezia species, as CFU clumps, glistening to dull and brittle and butyrous in texture, were occasionally formed. In those cases, plate counts were semiquantitative, thus resulting in higher inoculum size estimates ranging from 3.0 x 103 to 4.0 x 103 CFU/ml.
Microtiter plates were incubated at 32°C and were read at 48 h for M. furfur and M. pachydermatis and at 72 h for M. sympodialis, M. slooffiae, M. globosa, M. obtusa, M. restricta, and M. dermatis. For azoles, the MIC endpoint was defined as the lowest concentration which produced a predominant decrease (90% inhibition) of turbidity compared with that of the growth control. These stringent 90% endpoints were employed because, using reference strains and 20 of 53 haphazardly selected clinical strains in pilot experiments, it was established that intense trailing phenomena, expressed as persistent turbidity for drug concentrations above the MIC, were not observed. However, to allow for statistical discrepancies, an endpoint of 50% was employed when the number of Malassezia species tested was <10. For amphotericin B, the MIC endpoint was defined as the lowest concentration which completely inhibited growth (100% inhibition) because pilot BMD results had shown that trailing endpoints were not encountered. Results for M. furfur and M. pachydermatis tests were also read at 72 h to assess the extent to which trailing phenomena contribute to MIC endpoint variability.
Microtiter plates containing drug dilutions can be stored at 70°C for 6 months, as testing of QC strains showed that MIC readings were within the expected range.
(ii) Etest.
The agar formulations used for Etest (AB Biodisk, Solna, Sweden) comprised RPMI 1640 (Angus Buffers and Biochemicals) containing 1.5% agar (Difco, Detroit, Mich.) and 2% glucose buffered with MOPS (morpholinepropanesulfonic acid) and was prepared according to the manufacturer's instructions. The solid medium was supplemented with the aforementioned mixture of ox bile, glycerol, glycerol monostearate, and Tween 20, which was autoclaved at 121°C for 15 min, allowed to reach
50°C, and added to the sterile buffered RPMI 1640 agar suspension. The prepared plates can be stored at 6 to 8°C for not more than 2 weeks, as the medium dehydrates while the fatty acid supplements form exudates concentrated on the agar surface. Before being tested, each isolate was subcultured at least twice in LN agar (8) to ensure optimum growth.
For the Etest, 90-mm-diameter plates with 4-mm agar depth were used. The agar surface was inoculated using the Retro C80 inoculator (Biotools; AB Biodisk) with a nontoxic swab dipped in each suspension of Malassezia cells adjusted spectrophotometrically at 530 nm to a turbidity of 0.5 McFarland standard. The inoculated plates were left to stand for 15 min until the excess moisture was completely absorbed into the agar before the Etest strips (Nema C88; AB Biodisk) were applied to the agar surface of each plate. The plates were incubated at 32°C in a humid atmosphere, and MICs were recorded after 48 and 72 h. The MIC was taken as the lowest concentration at which the zone of complete inhibition intersected the strip. For the trailing Malassezia isolates, discernible (hazy) growth within the Etest inhibition ellipse was ignored.
For comparison purposes, Etest MIC assays were also performed using LN medium (8).
(iii) AM3 MIC determinations. To confirm that amphotericin B increased MICs, antibiotic medium 3 (AM3) broth and agar (Unipath, Ltd., Basingstoke, Hampshire, United Kingdom) supplemented with 2% dextrose was used for BMD and Etest assays, respectively. The medium was also supplemented with the same quantities of bile salt, fatty acid, and Tween 20 per liter. Incubation temperatures and reading of results were as described for the RPMI media.
(iv) Reference strains for QC. QC was performed throughout with M. furfur ATCC 14521 and M. globosa ATCC 96807. M. sympodialis ATCC 96803 was used as a reference strain. QC determinations were made on each test occasion and read at 48 h for M. furfur and at 72 h for M. globosa and M. sympodialis. The MICs on each test occasion were always within the same range.
(v) Analysis of results. Both on-scale and off-scale results were included in the analysis. The low off-scale MICs were left unchanged, and the high off-scale MICs were converted to the next-highest concentration. The data were log transformed before analysis and expressed as geometric means. Agreement was evaluated by concordance between the MICs determined by the two different susceptibility-testing methods. The BMD and Etest methods were considered to be in agreement when the MICs differed by no more than one twofold dilution. Intraclass correlation coefficients (ICCs) and 95% confidence intervals were calculated on log2-transformed data. The ICC was calculated to determine the variability of MICs and the reproducibility of each of the methods in the various subgroups. All values indicated average ICCs and showed absolute agreement with a maximum value of 1. All P values were two sided, and a P value of <0.05 was considered to indicate statistical significance. The statistical analysis was performed with Statistical Package for the Social Sciences (SPSS) software (version 10.0; SPSS Inc., Chicago, Ill.).
| RESULTS |
|---|
|
|
|---|
Susceptibility testing by BMD and Etest.
The fatty acid broth and solid RPMI medium allowed good Malassezia growth and MICs that were readily read after incubation at the optimum Malassezia incubation temperature (32°C) for 48 and 72 h. Reading times for microtiter plates and the Etest were species dependent. MICs of all drugs against M. furfur and M. pachydermatis could be recorded after 48 h, whereas MICs against all other Malassezia species were read after 72 h. For most M. furfur isolates and the M. pachydermatis strain, the difference between readings at 48 and 72 h was minimal (1 dilution). Certain M. furfur isolates (2 of 12) showed a rise in azole MICs by the BMD method between 48 and 72 h due to partial inhibition of growth over a range of antifungal concentrations (trailing phenomenon). On these rare occasions, persistent turbidity was encountered in wells 1 to 2 dilutions above the MIC endpoint at 72 h of incubation. Dispersion of clumps by pipetting made 48- and 72-h endpoint determinations reproducible. The formation of clumps, resulting from the butyrous-textured Malassezia colonies, was common in the course of testing the yeast, and it was also the reason for attaining semiquantitative, rather than quantitative, plate counts during estimations of the inoculum size. M. furfur clinical isolates also presented the trailing phenomenon when tested by the Etest (Fig. 1A). Trailing growth was recorded with Malassezia species (
4% of the isolates), including the QC type strain M. globosa ATCC 96807. No trailing was observed with terbinafine, which was tested only by the BMD method.
|
Azole MICs tested by Etest in LN medium could be obtained for M. furfur, M. slooffiae, M. dermatis, M. sympodialis, and M. pachydermatis, but as growth was slow, readings could only be recorded after 4 to 5 days (data not shown). Concordant amphotericin B MICs using RPMI 1640-fatty acid substrate and LN agar were not obtained, as MICs for the majority (85%) of strains showed low susceptibility (8 to
32 µg/ml) in LN medium. These high MICs could not be confirmed by AM3 (data not shown).
QC. Azole MICs, except fluconazole, in a series of 20 consecutive tests by BMD ranged from 0.03 to 0.125 µg/ml; terbinafine and amphotericin B MICs ranged from 0.03 to 0.25 µg/ml. Fluconazole MICs were typically higher than those of the other azoles and ranged between 4 and 16 µg/ml. Etest azole MICs, except that of fluconazole, in a series of 20 consecutive tests ranged from 0.03 to 0.25 µg/ml, and amphotericin B MICs ranged from 0.03 to 0.5 µg/ml. Etest fluconazole MICs ranged from 6 to 32 µg/ml. Only 1 out of 20 azole and terbinafine MICs was outside the recorded BMD and Etest range. No amphotericin B measurement was out of the control with either method.
Analysis of results. Overall, agreement within ±1 dilution between the BMD method and Etest was 84 to 97% for all drugs tested except amphotericin B (49%). ICCs were statistically very significant (P < 0.001) (Table 2). The correlation between levels of MICs obtained by the two methods is shown in Fig. 2 on log2-transformed data.
|
|
| DISCUSSION |
|---|
|
|
|---|
As previously recorded in Etest evaluation studies of other yeasts (15-17), the problem of trailing endpoints due to partial inhibition of growth by the azoles was also observed with Malassezia species. However, compared with Candida species, Malassezia yeasts displayed limited trailing phenomena after 48 h of incubation. In the BMD method, reading of MICs after 48 h of incubation was facilitated by dispersing the clumps formed in the one or two wells above the 48-h MICs. In addition, precise implementation of the specific criteria for reading Etest MICs, as explained in the Etest package insert and technical guide for yeasts (AB Biodisk), facilitates the reading of Malassezia endpoint MICs. In that respect, good agreement with BMD MICs was recorded when discernible growth within an established ellipse was ignored.
Although terbinafine, amorolfine, and established and newer azole compounds have been tested against Malassezia skin isolates by BMD methods (4, 12), data on the in vitro performance of amphotericin B are almost exclusively associated with cases of bloodstream infections (1, 9, 18). However, these studies were performed prior to the resolution of the taxonomic status of Malassezia species and the recognition of new species (5). Before 1996, Malassezia isolates were identified in a broad sense (sensu lato) and susceptibility testing was performed by a variety of methods in dissimilar media, while AM3 was not used before for confirmation of amphotericin B resistance. Therefore, the association between the amphotericin B resistance data presented here and the therapeutic outcomes reported in earlier studies is obscured.
In addition, there are limited data on amphotericin B susceptibility relating to a large sample size of Malassezia isolates from human skin, which, in immunocompromised hosts, constitutes the reservoir for bloodstream infections (1, 2, 6, 12, 21). Interestingly, of our clinical isolates, 3 of 12 M. furfur, 2 of 2 M. restricta, 7 of 22 M. globosa, and 4 of 4 M. slooffiae isolates showed high amphotericin B MICs by BMD and Etest, which in all cases were confirmed by testing in AM3. Of these strains, the single bloodstream isolate tested displayed AM3-confirmed resistance to amphotericin B. Although comparing MICs for skin and bloodstream isolates was beyond the aims of our study, this finding shows that data on the MIC ranges of Malassezia bloodstream isolates can be helpful in the clinical setting. The scarcity of recovery of Malassezia species from blood cultures has so far impaired antifungal susceptibility studies of a significant number of isolates and consequently correlation of amphotericin B in vitro response with the in vivo outcome. Despite that, the few published case reports give evidence for the poor therapeutic outcome of amphotericin B against Malassezia bloodstream infections. Amphotericin B is used empirically for management of Malassezia bloodstream infections, but with low response to therapy irrespective of the patient's immune status (1, 2, 10, 18, 19, 21). Therefore, testing the susceptibilities of these isolates by methods with established reproducibility may be indicated.
Despite the fact that growth of Malassezia species is well supported by the liquid and solid LN media, our study indicated that LN media may not be suitable for susceptibility testing of azoles. This was due to the long incubation time required before growth was adequate for MIC reading, thus enhancing trailing phenomena. Long incubation required for sharp MIC readings in LN media has been documented before (7), whereas the low growth rates of M. restricta, M. obtusa, and M. globosa recorded in LN-Alamar Blue broth (4) disqualifies the medium for azole MIC testing, at least for these three Malassezia species. In the present study, consistently higher amphotericin B MICs were recorded in LN media. The absence of MIC concordance between the LN and the fatty acid RPMI media and the subsequent unconfirmed resistance in AM3 indicates that the proposed RPMI media may be better candidates for amphotericin B MIC testing.
The fatty acid RPMI BMD and Etest susceptibility data reported here are not in agreement with those recorded using the urea BMD method (13), although they are not directly comparable. The utilization of a single metabolic marker as a growth indicator and the use of a large inoculum size (2.5 x 106 cells/ml), as opposed to an inoculum corresponding to 2.0 x 103 to 4.0 x 103 CFU/ml adopted in this study, may have accounted for the low susceptibilities to itraconazole and terbinafine of the seven Malassezia species tested in urea broth.
Our data indicated low agreement (49%) between the BMD method and Etest concerning amphotericin B. This is attributed to the consistently higher Etest MICs (1 to 2 dilutions) obtained for strains with BMD MICs of
4 µg/ml, thus influencing the agreement between the two methods. In real terms, this does not appear to constitute significant divergence of Etest MICs from those obtained by BMD, because all higher amphotericin B Etest MICs were obtained only for strains with high BMD MICs. This indicates that the low amphotericin B susceptibilities of certain Malassezia strains reflect a real phenotypic trait, which is possibly enhanced by testing it on solid fatty acid RPMI medium. Moreover, the observed low amphotericin B M. restricta and M. slooffiae susceptibilities may be representing a random event rather than a medium- or species-associated phenomenon, due to the small number of strains tested.
Overall, the agreement of BMD and Etest under the stringent conditions of ±1 dilution was 88%. This provides confidence that reproducible MICs can be obtained by using fatty acid RPMI for BMD and Etest. Both methods appear reliable and have good correlation (r = 0.89), with a highly significant intraclass correlation coefficient (P < 0.001). Nonetheless, the clinical relevance of MICs for the eight Malassezia species tested, and MICs for new species (23) yet to be determined, remains to be ascertained.
| ACKNOWLEDGMENTS |
|---|
| FOOTNOTES |
|---|
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Antimicrob. Agents Chemother. | Clin. Microbiol. Rev. |
|---|---|
| Clin. Vaccine Immunol. | ALL ASM JOURNALS |
|---|