Available online at journal homepage:

Transcription

1 Kaohsiung Journal of Medical Sciences (2012) 28, 306e315 Available online at journal homepage: ORIGINAL ARTICLE Fluconazole exposure rather than clonal spreading is correlated with the emergence of Candida glabrata with cross-resistance to triazole antifungal agents Tun-Chieh Chen a,d, Yen-Hsu Chen b,d,e, Yee-Chun Chen f,g,h, Po-Liang Lu b,c,d, * a Department of Internal Medicine, Kaohsiung Municipal Ta-Tung Hospital, Kaohsiung Medical University, Kaohsiung City, Taiwan b Division of Infectious Diseases, Department of Internal Medicine, Kaohsiung Medical University Hospital, Kaohsiung Medical University, Kaohsiung City, Taiwan c Faculty of Medicine, College of Medicine, Kaohsiung Medical University, Kaohsiung City, Taiwan d Graduate Institute of Medicine, College of Medicine,Kaohsiung Medical University, Kaohsiung City, Taiwan e Tropical Medicine Research Center, College of Medicine, Kaohsiung Medical University, Kaohsiung City, Taiwan f Department of Internal Medicine, National Taiwan University Hospital, Taipei City, Taiwan g Center for Infection Control, National Taiwan University Hospital, Taipei City, Taiwan h Department of Medicine, College of Medicine, National Taiwan University, Taipei City, Taiwan Received 6 April 2011; accepted 17 August 2011 Available online 4 April 2012 KEYWORDS Cross-resistance; Candida glabrata; Triazole Abstract The emergence of antifungal resistance in Candida species has raised concern in recent years, especially resistance toward triazole. Several newer triazole antifungal agents have been introduced which have a broader spectrum for fungal infections, such as voriconazole. However, cross-resistance among triazoles is a major concern with regard to their clinical application. Antifungal susceptibility was performed using E-test for 166 clinical isolates (29 blood and 137 nonblood isolates) in 2003 and We applied pulsed-field gel electrophoresis for genotyping. Ninety isolates of C. albicans, 47 isolates of C. tropicalis, 27 isolates of C. glabrata, and two isolates of C. krusei were included. All isolates were susceptible to amphotericin B. Eleven (40.7%) of the 27 C. glabrata had intermediate resistance to caspofungin. Fortyseven (28.3%) of the 166 isolates were not susceptible to fluconazole, including two C. albicans, 16C. tropicalis, 27C. glabrata, andtwoc. krusei isolates. All except seven of the C. glabrata isolates were susceptible to voriconazole. All the triazole drugs had a positive correlation among their minimum inhibitory concentrations (MICs). Fluconazole MIC was a good predictor for susceptibility to voriconazole, as determined using a receiver operating * Corresponding author. Number 100, Tzyou 1st Road 807, Kaohsiung City, Taiwan. address: d830166@cc.kmu.edu.tw (P.-L. Lu) X/$36 Copyright ª 2012, Elsevier Taiwan LLC. All rights reserved. doi: /j.kjms

2 Fluconazole use and triazole cross-resistance 307 characteristic curve. Furthermore, a high diversity of pulsotypes for the 27 clinical isolates of C. glabrata was observed. Previous fluconazole exposure within 3 months was associated with reduced triazole susceptibility for C. glabrata. We demonstrated a significant positive correlation of MIC values among the four tested triazole drugs. No amphotericin B and caspofungin resistant isolates were found in this study. The cross-resistance to triazole among C. glabrata isolates was associated with previous fluconazole exposure as opposed to clonal spreading. Selection pressure due to fluconazole use may play a major role in triazole cross-resistance. Copyright ª 2012, Elsevier Taiwan LLC. All rights reserved. Introduction Candida infections have become more prevalent in recent decades because of the increasing number of immunocompromised hosts and critically ill patients [1,2]. In Taiwan, Candida species were the first and the second most common pathogens of nosocomial infections in medical centers and regional hospitals, respectively, in 2008 [3]. Fluconazole has been the most widely used antifungal agent for invasive candidiasis since the 1990s, and concern over the emergence of fluconazole resistance has risen, especially with the increasing use of fluconazole for the treatment and prophylaxis of candidal infections [2]. A nationwide survey across Taiwan showed increasing fluconazole resistance rates, from 1.9% in 2002 to 17.1% in 2006, especially among nonblood or nonalbicans isolates [4]. Furthermore, co-resistance to both amphotericin B and fluconazole has been observed, indicating a serious problem for the treatment of these infections [4]. The increasing seriousness of Candida infections within hospitals with regard to the increasing incidence and resistance rates in Taiwan is reflected worldwide [2]. However, the threat of fluconazole resistance rates in C. glabrata in southern Taiwan is particularly serious, with two nationwide surveys reporting 40e81% of isolates with nonsusceptibility [5,6]. Although antifungal susceptibility testing is not routinely performed, local reference data for antifungal susceptibility and the epidemiology of species distribution, the existence of occult outbreaks of resistant strains, and the analysis of risk factors relevant to resistance are important. Antimicrobial resistance originates from either antimicrobial selection pressure or a resistant gene having spread clonally. Here we report the analysis of 166 clinical isolates from a Taiwan medical center and compare the data of candidemia isolates from 1998 [7] to elucidate the change in antifungal resistance and its possible mechanisms. Material and methods Study duration and population All Candida species for this study were consecutively isolated from clinical specimens of Kaohsiung Medical University Hospital between September 1, 2003, and May 31, 2004, and sputum specimens that were not confirmed to be Candida pneumonia were excluded. Only the first Candida isolate of multiple isolates of the same species and from the same patient was included for further study. Chart review was performed to collect patient demographic information, including medical history (malignancy, total parenteral nutrition, neutropenia, intensive care unit admission, diabetes mellitus, hypertension, and any underlying diseases), invasive procedures (including use of ventilator, urinary catheters, among others), use of immunosuppression drugs and antifungal agents (drugs for more than 1 day, within 3 months), and laboratory data using a standardized case record form. Clinical Candida isolates were identified to the species level by colonial characteristics, microscopic features, carbohydrate assimilation-fermentation studies, and germ tube tests, and then confirmed using the ID 32 C system (biomerieux, Hazelwood, MO, USA). Antifungal susceptibility testing was performed and minimum inhibitory concentrations (MICs) were determined by the E-test method (AB Biodisk, Solna, Sweden) for amphotericin B, fluconazole, itraconazole, ketoconazole, voriconazole, and caspofungin. RPMI-1640 agar buffered with 4-morpholinepropanesulfonic acid (MOPS, ph 7.0) was prepared for the E-tests. A total of 0.5 McFarland standard inocula were applied to the RPMI agar with a cotton swab. The MIC endpoints were read after 48 hours of incubation at 35 C [8e10]. C. albicans ATCC 90028, C. krusei ATCC 6258, and C. parapsilosis ATCC were used as control strains. Results for susceptibility were considered eligible only when the MICs of the control strains were all within the standard range [11]. We used the recently revised the Clinical and Laboratory Standards Institute (CLSI) clinical breakpoints to define susceptibility categories of Candida isolates to caspofungin and fluconazole [12,13]. Caspofungin MIC values of S1 mg/ml were considered resistant for C. albicans, C. tropicalis, and C. krusei, and MIC results of S8 mg/ml were categorized as resistant for C. parapsilosis; caspofungin MIC values of S0.5 mg/ml were considered resistant for C. glabrata. Fluconazole MIC results of S8 mg/ml were defined as resistant for C. albicans, C. parapsilosis, and C. tropicalis, and MICs S64 mg/ml were considered resistant for C. glabrata. All isolates of C. krusei were defined as resistant to fluconazole. Interpretative criteria of susceptibility for Candida species to itraconazole and voriconazole were in accordance with the CLSI guidelines in 2008 [14]. Isolates with an itraconazole MIC of mg/ ml were classified as S, those with an itraconazole MIC of >0.5 mg/ml were classified as resistance (R), and those in between were classified as susceptible-dose dependent (S- DD). Susceptibility to voriconazole was defined as a voriconazole MIC 1 mg/ml. Amphotericin B-resistant Candida isolates were defined as those with an amphotericin B MIC > 1 mg/ml, as previously proposed [15]. Ketoconazoleresistant Candida isolates were defined as those with a ketoconazole MIC > mg/ml [16].

3 Table 1 Species (isolates) MICs and antifungal susceptibility of Candida species. Amphotericin B Fluconazole Itraconazole Ketoconazole Voriconazole Caspofungin Range S/R Range S/SDD/R a S/SDD/R b Range S/SDD/R Range S/R Range S/SDD/R Range S/R a S/I/R b S% S% a S% b S% S% S% S% a S% b C. albicans Blood (21) 0.004e / e >256 20/0/1 20/0/ e >32 20/0/ e >32 20/ e /0/ e /0 21/0/ Urine (52) 0.008e / e >256 51/0/1 51/0/ e >32 47/4/ e >32 51/ e /0/ e /0 52/0/ Other (17) 0.012e /0 0.38e1.5 17/0/0 17/0/ e /1/ e / e /0/ e /0 17/0/ Subtotal (90) 0.004e / e >256 88/0/2 88/0/ e >32 83/5/ e >32 88/ e /0/ e /0 90/0/ C. tropicalis Blood (7) 0.016e0.5 7/0 1e8 7/0/0 6/0/ e1 4/1/ e / e0.38 7/0/ e /0 7/0/ Urine (35) 0.012e0.5 35/0 0.25e12 33/2/0 23/7/ e2 14/8/ e / e0.5 35/0/ e /0 35/0/ Other (5) 0.016e0.5 5/0 1e8 5/0/0 2/2/ e3 1/1/ e / e0.38 5/0/ e /0 5/0/ Subtotal (47) 0.012e0.5 47/0 0.25e12 45/2/0 31/9/ e3 19/10/ e / e0.5 47/0/ e /0 47/0/ C. glabrata Blood (1) / /0/0 0/1/ /0/ / /0/ /0 0/1/ Urine (20) 0.023e /0 2e256 8/4/8 0/12/ e32 2/3/ e6 7/ e6 13/0/ e /0 15/5/ Other (6) 0.023e0.5 6/0 3e12 4/2/0 0/6/ e3 1/1/ e0.25 2/ e0.38 6/0/ e0.47 6/0 1/5/ Subtotal (27) 0.023e0.5 27/0 2e256 13/6/8 0/19/ e32 4/4/ e6 10/ e6 20/0/ e /0 16/11/ C. krusei Blood (0) d d d d d d d d d d d d d d d d d d d d d d d d d d d Urine (2) 0.047e0.75 2/0 64 e >256 0/0/2 0/0/2 1.5e4 0/0/2 0.5e0.75 0/2 0.25e0.38 2/0/0 0.19e0.25 2/0 2/0/ Other (0) d d d d d d d d d d d d d d d d d d d d d d d 308 T.-C. Chen et al.

4 Fluconazole use and triazole cross-resistance 309 Subtotal (2) 0.047e0.75 2/0 64 e >256 0/0/2 0/0/2 1.5e4 0/0/2 0.5e0.75 0/2 0.25e0.38 2/0/0 0.19e0.25 2/0 2/0/ Total (166) 0.004e / e > /8/12 119/28/ e >32 106/19/ e >32 145/ e6 159/0/ e /0 155/11/ CLSI Z Clinical and Laboratory Standards Institute; MIC Z minimum inhibitory concentration; R Z resistance; S Z susceptible; SDD Z susceptible-dose dependent. a Antifungal susceptibility test interpretive category by CLSI M27-A3 (2008). b Antifungal susceptibility test interpretive category by updated CLSI criteria (2011). Genotyping using pulsed field gel electrophoresis Pulsed field gel electrophoresis (PFGE) for C. glabrata utilized the restriction enzymes SfiI and BssHII, as previously described [17]. To identify PFGE polymorphisms, each sample was analyzed using Molecular Analyst Fingerprinting, Fingerprinting Plus, and Fingerprinting DST software (Bio-Rad Laboratories, Richmond, CA, USA). The grouping method was performed to produce a dendrogram from the matrix via the unweighted pair group method using the arithmetic average clustering technique after calculating similarities using Pearson correlation coefficient between every pair of organisms. Statistical analysis All data were analyzed using SPSS version 17.0 (SPSS Inc., Chicago, IL, USA). Categorical variables were compared using univariate analysis with the Chi-square test. Continuous variables were analyzed with the student t test or one-way analysis of variance (ANOVA). A p value of <0.05 was considered statistically significant (two-tailed). Multivariate analyses were conducted using a logistic regression model for the variables that were significant in the univariate analysis. Pairwise correlations of the logarithm of six antifungal MICs to the base 2 (Log2) were performed using JMP version 7.0 (SAS Institute, NC, USA). A receiver operating characteristic (ROC) curve was used to determine the optimal cutoff point of MICs for fluconazole, itraconazole, and ketoconazole to predict voriconazole nonsusceptibility, which was performed using MedCalc version (MedCalc, Mariakerke, Belgium). Results From September 1, 2003, through May 31, 2004, a total of 166 clinical isolates (29 blood and 137 nonblood isolates) were collected. There were 90 isolates (90/166, 54.2%) of C. albicans, 47 of C. tropicalis, 27 of C. glabrata, and two of C. krusei. The susceptibility to antifungal agents of the various species and specimen origins is shown in Table 1. All isolates were susceptible to amphotericin B. Eleven (40.7%) of the 27 C. glabrata showed intermediate resistance to caspofungin. Forty-seven (28.3%) of the 166 isolates were not susceptible to fluconazole, including two C. albicans, 16 C. tropicalis, 27 C. glabrata, and two C. krusei isolates. All isolates were susceptible to voriconazole except for seven isolates of C. glabrata obtained from urine specimens. The co-resistance relationships of the six antifungal agents were examined by paired correlation of Log2 values of MICs (Fig. 1). All the triazole drugs (fluconazole, ketoconazole, itraconazole, and voriconazole) had MIC values that demonstrated a high correlation to each other in that all combinations had correlation coefficients >0.74. The susceptibility of the older triazoles (fluconazole, itraconazole, and ketoconazole) also had a significant linear correlation to voriconazole against C. glabrata (p values of Spearman correlation: < for fluconazole, 0.02 for itraconazole, and 0.01 for ketoconazole). However, only the susceptibility of fluconazole could significantly predict

5 310 T.-C. Chen et al. Figure 1. Scatterplot matrix and pairwise correlation of log 2 values for the minimum inhibitory concentrations of the six tested antifungal agents. the susceptibility of C. glabrata for voriconazole (p < 0.001) (Table 2). An ROC curve was used to determine the cut-off MIC values of fluconazole, itraconazole, or ketoconazole for nonsusceptibility of C. glabrata and all Candida species to voriconazole. Significant high areas under the curve and likelihood ratios are shown in Fig. 2A, when MICs of fluconazole >16 mg/ml, itraconazole >3 mg/ ml, and ketoconazole >1.5 mg/ml to predict C. glabrata nonsusceptibility to voriconazole. Significant high areas under the curve and likelihood ratios are also shown in Fig. 2B, when MICs of fluconazole >64 mg/ml, itraconazole >4 mg/ml, and ketoconazole >1.5 mg/ml to predict voriconazole nonsuscetibility for all Candida species. There was no linear relationship among amphotericin B, caspofungin, and the four tested azole drugs. There was also no significant relationship between amphotericin B and caspofungin (correlation coefficient: , p Z 0.883). With regard to the typing result of the 27 C. glabrata isolates, four (14.8%) isolates that were susceptible to both voriconazole and fluconazole were of the same pulsotype. Table 2 The correlation of susceptibility to different triazoles and the prediction of voriconazole susceptibility. Isolates Fluconazole Itraconazole Ketoconazole All Candida species S SDD R S SDD R S R Voriconazole susceptible 146 a 8 a 5 a b 28 b 12 b Voriconazole nonsusceptible Chi-square test p value <0.001 a,b <0.001 <0.001 Spearman correlation p value <0.001 a,b <0.001 <0.001 Candida glabrata S SDD R S SDD R S R Voriconazole susceptible isolates 13 a 6 a 1 a b 19 b 1 b Voriconazole nonsusceptible isolates Chi-square test p value <0.001 a,b Spearman correlation p value <0.001 a,b CLSI Z Clinical and Laboratory Standards Institute; MIC Z minimum inhibitory concentration; R Z resistance; S Z susceptible; SDD Z susceptible-dose dependent. a Antifungal susceptibility test interpretive category by CLSI M27-A3 (2008). b Antifungal susceptibility test interpretive category by updated CLSI criteria (2011).

6 Fluconazole use and triazole cross-resistance 311 Figure 2. (A) Using the ROC curve method to determine cut-off MIC values of surrogate markersdfluconazole, itraconazole, or ketoconazoledfor voriconazole nonsusceptibility of C. glabrata; (B) using the ROC curve method to determine cut-off MIC values of surrogate markersdfluconazole, itraconazole, or ketoconazoledfor voriconazole nonsusceptibility of the Candida species. MIC Z minimum inhibitory concentration; ROC Z receiver operating characteristic. The majority of the isolates, however, were not of identical pulsotypes (Fig. 3). This indicated that there was no clonal transmission of voriconazole-resistant C. glabrata. Excluding 12 cases whose clinical data were not adequately collected, the clinical characteristics of 154 cases were further correlated with species and susceptibility differences. There was no correlation between Candida species and specimen types. No factor of demographics, underlying disease, or invasive procedure was significantly related with a specific Candida species (we excluded C. krusei from the analysis due to its small case number). Only fluconazole exposure in the last 3 months was significantly associated with geometric means of MIC for fluconazole, itraconazole, and ketoconazole (p Z 0.012, 0.006, and 0.028, respectively) against C. glabrata, and fluconazole exposure in the last 3 months was also marginally associated with the MIC for voriconazole (p Z 0.051). There was also a significant association between fluconazole exposure and antifungal resistance rates against fluconazole (p Z 0.023) and ketoconazole (p Z 0.021), and there was a borderline association between fluconazole exposure and voriconazole resistance [p Z (Table 3)]. No association was observed between fluconazole resistance and the use of nontriazole antifungal agents, such as amphotericin B or nystatin, and the duration of fluconazole exposure. Other factors including demographics, underlying diseases, and invasive procedures that were related to resistance toward antifungal agents, including triazoles, amphotericin B, and caspofungin, were not identified. Discussion The current study revealed that most Candida isolates (71.7%) were still susceptible to fluconazole, except for C. glabrata isolated from urine and other nonblood specimens (100% were nonsusceptible). All Candida isolates demonstrated 100% susceptibility to amphotericin B and caspofungin, except 40.7% of the 27 C. glabrata, which showed intermediate resistance to caspofungin. Voriconazole, a new triazole, also had better activity against all Candida isolates relative to fluconazole, except for seven urine isolates that were resistant to voriconazole. A review of invasive candidiasis in Taiwan [2] and a recent nationwide survey [4] also revealed little variation in the species distribution of clinical Candida isolates based on geographic and secular differences, and around one-half of the Candida isolates were C. albicans, which was similar to the results of this study. These results differ from those of studies in North America [18] and Ireland [19], in which the prevalence of nonalbicans Candida was higher, especially in patients who had cancer and among those who were critically ill. This was probably related to the common application of fluconazole prophylaxis. A major secular change in fluconazole susceptibility was noted as a shift from susceptible to susceptible-dose dependent for C. glabrata isolates despite the stable fluconazole resistant rates demonstrated in several studies [2,4,6]. A previous survey reported that Candida isolates in southern Taiwan had a higher fluconazole resistance rate (11.8%) compared with those in northern and central Taiwan (2.5% and 6.5%) [20].

7 312 T.-C. Chen et al. Figure 3. Dendrogram of pulsed field gel electrophoresis results for 27 C. glabrata isolates. However, the method used to collect isolates in that study (10 C. albicans and 40 non-albicans from each hospital) did not reflect the species distribution; therefore, this might have influenced the drug resistance rate because different species of Candida have different resistance tendencies to various drugs. Each hospital in Taiwan had a different resistance rates to fluconazole, which ranged from 0% to 24% [21]. C. glabrata isolates in southern Taiwan were found to have a higher fluconazole nonsusceptible rate (40%) than isolates in northern (29%) and central (14%) regions [5]. Our study, conducted in a hospital in southern Taiwan, showed a serious threat of triazole resistance for C. glabrata. Clinicians should there be alert to the link between triazole resistance and this species of Candida when prescribing antifungal agents. Since routine antifungal susceptibility testing is not recommended [22], survey data of resistance rates for each species are necessary to assist the clinician in choosing antifungal agents. We compared the susceptibility pattern reported in 1997 [7] in our hospital to the data in this current study (Table 4). A lower susceptibility rate (64.7%) of C. tropicalis blood isolates for amphotericin B was found in 1997 compared with the 100% susceptible rate observed in this study. A low susceptibility rate (48.1%) of C. glabrata isolates to fluconazole was a significant finding when compared with the 1997 data. The observations that 35% of C. glabrata isolates from urine specimens were resistant to voriconazole and the cross-resistance phenomenon among triazoles make amphotericin B and caspofungin choices when triazole treatment fails when fluconazole was suggested by guideline as the first-line treatment for candiduria [23]. Although echinocandins were proven useful for Table 3 The relationship of fluconazole exposure and geometric means of MICs and resistance percentages in different Candida species. Geometric means of MIC p value Resistance (%) p value Fluconazole exposure Yes No Yes No C. albicans Numbers: (86) (47) (39) Fluconazole a % 0% Itraconazole % 0% Ketoconazole % 0% Voriconazole % 0% d C. tropicalis Numbers: (41) (27) (14) Fluconazole a % 14.29% Itraconazole % 11.11% Ketoconazole % 0% d Voriconazole % 0% d C. glabrata Numbers: (26) (11) (15) Fluconazole a % 7.14% Itraconazole % 46.67% Ketoconazole % 6.67% Voriconazole % 6.67% CLSI Z Clinical and Laboratory Standards Institute; MIC Z minimum inhibitory concentration. a Antifungal susceptibility test interpretive category by updated CLSI criteria (2011).

8 Fluconazole use and triazole cross-resistance 313 Table 4 Comparison of antifungal susceptibility data of isolates in 1997 and from 2003 to Amphotericin B Fluconazole a b Number MIC50 MIC 90 Range S% R% MIC 50 MIC 90 Range S% S-DD % R% C. albicans A e e > (95.2) 0 (0) 2.2 (4.8) 100 (100) 0 (0) 0.75 (0.5) 1 (1) e >256 (0.38 e >256) 0.004e0.25 (0.004e0.25) 0.25 (0.011) B. 90 (21) c (0.023) C. tropicalis A e e (100) 0 (0) 1.5 (1) 8 (8) 0.25e12 (1-8) 95.7 (100) 4.3 (0) 0 (0) 0.012e0.5 (0.016e0.5) 0.5 (0.5) B. 47 (7) c (0.023) C. glabrata A e e B. 27 (1) c e (100) 0 (0) e256 (3) 48.1 (100) 2.2 (0) 29.6 (0) (0.064) CLSI Z Clinical and Laboratory Standards Institute; MIC Z minimum inhibitory concentration; S-DD Z susceptible-dose dependent. a Antifungal susceptibility test interpretive category by CLSI M27-A3 (2008). b A Z the data for candidemia isolates in 1998 [7]; BZ Data for isolates from 2003 to 2004 (from the current study). c Numbers and percentages in parenthesis represent the data from blood isolates in the current study. treating triazole cross-resistant C. glabrata infections [24], echinocandins have low urine concentrations and are not recommended for treating urinary tract infections. As such, nephrotoxic agents, such as amphotericin and flucytosine, should be used for fluconazole-resistant candiduria. The possible role of drug efflux pumps, including CgCDR1/2 and MDR1, induced by triazole administration has been demonstrated in clinical cases [25e27] and in vitro studies [28e30]. The high diversity of pulsotypes and lack of clonal spreading of triazole-resistant C. glabrata in the hospital indicated that the high resistant rates to triazoles among the C. glabrata isolates was not due to an occult outbreak within the hospital. C. glabrata was found to be related with previous antifungal use in epidemiological analysis and a higher resistance rate to triazoles indicated the existence of multiple triazole-resistant clones within the hospital setting. Three studies also supported the observation that previous fluconazole exposure is associated with the development of fluconazole-resistant C. glabrata fungemia [31e33]. Our results are different from three previous local studies that have reported specific clones with fluconazole resistance were circulating in hospitals in Taiwan [34e36]. Similar to the previously reported strong positive correlation of fluconazole and voriconazole MICs [37], here we report that the MICs of the four triazole drugs were highly correlated. With regard to the four triazole drugs, the lowest rate of susceptibility was to itraconazole (63.9%), with susceptible rates of 87.3% to ketoconazole, 88% to fluconazole, and 95.8% to voriconazole. It is noteworthy that these isolates were collected before the introduction of voriconazole use in the hospital, so there was no selection pressure for resistance to voriconazole. The presence of voriconazole resistance, though at a low rate, may have originated from the cross-resistance towards other triazoles used. Fluconazole has been reported to be a surrogate marker for susceptibility to ravuconazole [38], voriconazole [39], and posaconazole [40]. We compared these three older triazoles, and all demonstrated good correlation of their MICs to voriconazole. However, only the CLSI breakpoints of fluconazole susceptibility had a significant correlation to the susceptibility to voriconazole. Therefore, susceptibility and resistance to fluconazole can be a potent predictor for susceptibility to voriconazole. No amphotericin B resistant or caspofungin resistant isolate was found in this hospital survey, indicating the effectiveness of the two drugs; however, there is an alert for amphotericin B and fluconazole resistant isolates in Taiwan [20]. There were some limitations in the current study. First, the Candida isolates we analyzed were collected between 2003 and The Candida species distribution and antifungal susceptibility may have changed since then. However, we used the isolates from 2003 to 2004 because we wanted to analyze the antifungal susceptibility and epidemiology before introducing voriconazole in Taiwan to exclude the impact of voriconazole itself. Second, the MIC values in the current study were determined by E-test method instead of the standard broth dilution method in the CLSI guideline. Nevertheless, the results by E-test showed good correlation with those by broth dilution test [8e10]. In conclusion, C. albicans was the most common species, and most Candida isolates in the current study were

9 314 T.-C. Chen et al. susceptible to fluconazole, except for the emergence of triazole cross-resistant C. glabrata from urine specimens. The cross-resistant C. glabrata were associated with previous fluconazole exposure as opposed to clonal spreading. Fluconazole was a good predictor for susceptibility to voriconazole, and indeed better than itraconazole or ketoconazole. Amphotericin B and caspofungin retained excellent activity toward all Candida species, even toward triazole-resistant C. glabrata. Clinicians should be cautious of triazole cross-resistance, especially for patients with prior fluconazole prophylaxis or treatment and for patients with a risk for C. glabrata infections. According to the new clinical practice guidelines for the management of candidiasis by the Infectious Diseases Society of America [23], echinocandins or lipid-formulation amphotericin B should be considered in these cases. Acknowledgments This work was supported by Kaohsiung Medical University Hospital (93-KMUH-010), Taiwan. References [1] Chen YC, Chang SC, Sun CC, Yang LS, Hsieh WC, Luh KT. Secular trends in the epidemiology of noscomial fungal infections at a teaching hospital in Taiwan, 1981 to Infect Control Hosp Epidemiol 1997;18:369e75. [2] Ruan SY, Hsueh PR. Invasive candidiasis: an overview from Taiwan. J Formos Med Assoc 2009;108:443e51. [3] Centers for Disease Control, R.O.C. (Taiwan). Statistics of Communicable Diseases and Surveillance Report. R.O.C. (Taiwan): Taipei City: Centers for Disease Control; e44. [4] Yang YL, Wang AH, Wang CW, Cheng WT, Li SY, Lo HJ, et al. Susceptibilities to amphotericin B and fluconazole of Candida species in Taiwan surveillance of antimicrobial resistance of yeasts Diag Microbiol Infect Dis 2008;61:175e80. [5] Hsueh PR, Lau YJ, Chuang YC, Wan JH, Huang WK, Shyr JM, et al. Antifungal susceptibilities of clinical isolates of Candida species, Cryptococcus neoformans, and Aspergillus species from Taiwan: surveillance of multicenter antimicrobial resistance in Taiwan program data from Antimicrob Agents Chemother 2005;49:512e7. [6] Ruan SY, Chu CC, Hsueh PR. In vitro susceptibilities of invasive isolates of Candida species: rapid increase in rates of fluconazole susceptible-dose dependent Candida glabrata isolates. Antimicrob Agents Chemother 2008;52:2919e22. [7] Chen TC, Chen YH, Tsai JJ, Peng CF, Lu PL, Chang K, et al. Epidemiologic analysis and antifungal susceptibility of Candida blood isolates in southern Taiwan. J Microbiol Immunol Infect 2005;38:200e10. [8] Wanger A, Mills K, Nelson PW, Rex JH. Comparison of Etest and National committee for clinical laboratory standards broth macrodilution method for antifungal susceptibility testing: enhanced ability to detect amphotericin B-resistant Candida isolates. Antimicrob Agents Chemother 1995;39:2520e2. [9] Pfaller MA, Messer SA, Karlsson A, Bolmström A. Evaluation of the Etest method for determining fluconazole susceptibilities of 402 clinical yeast isolates by using three different agar media. J Clin Microbiol 1998;36:2586e9. [10] Pfaller MA, Messer SA, Houston A, Mills K, Bolmstrom A, Jones RN. Evaluation of the Etest method for determining voriconazole susceptibilities of 312 clinical isolates of Candida species by using three different agar media. J Clin Microbiol 2000;38:3715e7. [11] National Committee for Clinical Laboratory Standards. Reference Method for Broth Dilution Antifungal Susceptibility Testing of Yeasts. Approved standard. 2nd ed. Wayne, PA: National Committee for Clinical Laboratory Standards; [12] Pfaller MA, Andes D, Diekema DJ, Espinel-Ingroff A, Sheehan D. CLSI subcommittee for antifungal susceptibility testing. Wild-type MIC distributions, epidemiological cutoff values and species-specific clinical breakpoints for fluconazole and Candida: time for harmonization of CLSI and EUCAST broth microdilution methods. Drug Resist Updat 2010;13: 180e95. [13] Pfaller MA, Diekema DJ, Andes D, Arendrup MC, Brown SD, Lockhart SR, et al. Clinical breakpoints for the echinocandins and Candida revisited: Integration of molecular, clinical, and microbiological data to arrive at species-specific interpretive criteria. Drug Resist Updat 2011;14:164e76. [14] National Committee for Clinical Laboratory Standards. Reference Method for Broth Dilution Antifungal Susceptibility Testing of Yeasts. 3rd ed. Wayne, PA: National Committee for Clinical Laboratory Standards; Approved standard. [15] Nguyen MH, Clancy CJ, Yu VL, Yu YC, Morris AJ, Snydman DR, et al. Do in vitro susceptibility data predict the microbiologic response to amphotericin B? Results of a prospective study of patients with Candida fungemia. J Infect Dis 1998;177:425e30. [16] Rex JH, Pfaller MA, Walsh TJ, Chaturvedi V, Espinel-Ingroff A, Ghannoum MA, et al. Antifungal susceptibility testing: practical aspects and current challenges. Clin Microbiol Rev 2001; 14:643e58. [17] Lu PL, Hsueh PR, Chen YC, Teng LJ, Chang SC, Luh KT. Clones of Lactobacillus casei and Torulopsis glabrata associated with recurrent abdominal wall abscess. J Formos Med Assoc 1999; 98:356e60. [18] Trick WE, Fridkin SK, Edwards JR, Hajjeh RA, Gaynes RP. National Nosocomial Infections Surveillance System Hospitals. Secular trend of hospital-acquired candidemia among intensive care unit patients in the United States during Clin Infect Dis 2002;35:627e30. [19] McMullan R, McClurg R, Xu J, Moore JE, Millar BC, Crowe M, et al. Trends in the epidemiology of Candida bloodstream infections in Northern Ireland between january 1984 and december J Infect 2002;45:25e8. [20] Yang YL, Ho YA, Cheng HH, Lo HJ. TSARY Hospitals. Distribution and susceptibility to amphotericin B and fluconazole of Candida spp. isolated from Taiwan. Epidemiol Infect 2005; 133:325e30. [21] Yang YL, Cheng HH, Ho YA, Hsiao CF, Lo HJ. Fluconazole resistance rate of Candida species from different regions and hospital types in Taiwan. J Microbiol Immunol Infect 2003;36: 187e91. [22] Klepser ME. Antifungal resistance among Candida species. Pharmacotherapy 2001;21:124Se32S. [23] Pappas PG, Kauffman CA, Andes D, Benjamin Jr DK, Calandra TF, Edwards Jr JE, et al. Clinical practice guidelines for the management of candidiasis: 2009 update by the infectious diseases society of America. Clin Infect Dis 2009;48: 503e35. [24] Posteraro B, Sanguinetti M, Fiori B, La Sorda M, Spanu T, Sanglard D, et al. Caspofungin activity against clinical isolates of azole cross-resistant Candida glabrata overexpressing efflux pump genes. J Antimicrob Chemother 2006;58:458e61. [25] Magill SS, Shields C, Sears CL, Choti M, Merz WG. Triazole cross-resistance among Candida spp.: case report, occurrence among bloodstream isolates, and implications for antifungal therapy. J Clin Microbiol 2006;44:529e35. [26] Panackal AA, Gribskov JL, Staab JF, Kirby KA, Rinaldi M, Marr KA. Clinical significance of azole antifungal drug cross-

10 Fluconazole use and triazole cross-resistance 315 resistance in Candida glabrata. J Clin Microbiol 2006;44: 1740e3. [27] Posteraro B, Tumbarello M, La Sorda M, Spanu T, Trecarichi EM, De Bernardis F, et al. Azole resistance of Candida glabrata in a case of recurrent fungemia. J Clin Microbiol 2006;44:3046e7. [28] Sanguinetti M, Posteraro B, Fiori B, Ranno S, Torelli R, Fadda G. Mechanisms of azole resistance in clinical isolates of Candida glabrata collected during a hospital survey of antifungal resistance. Antimicrob Agents Chemother 2005;49: 668e79. [29] Borst A, Raimer MT, Warnock DW, Morrison CJ, Arthington- Skaggs BA. Rapid acquisition of stable azole resistance by Candida glabrata isolates obtained before the clinical introduction of fluconazole. Antimicrob Agents Chemother 2005; 49:783e7. [30] Cernicka J, Subik J. Resistance mechanisms in fluconazoleresistant Candida albicans isolates from vaginal candidiasis. Int J Antimicrob Agents 2006;27:403e8. [31] Tumbarello M, Sanguinetti M, Trecarichi EM, La Sorda M, Rossi M, de Carolis E, et al. Fungemia caused by Candida glabrata with reduced susceptibility to fluconazole due to altered gene expression: risk factors, antifungal treatment and outcome. J Antimicrob Chemother 2008;62:1379e85. [32] Lee I, Fishman NO, Zaoutis TE, Morales KH, Weiner MG, Synnestvedt M, et al. Risk factors Risk factors for fluconazoleresistant Candida glabrata bloodstream infections. Arch Intern Med 2009;169:379e83. [33] Lee I, Zaoutis TE, Fishman NO, Morales KH, Nachamkin I, Lautenbach E. Risk factors for fluconazole resistance in patients with Candida glabrata bloodstream infection: potential impact on control group selection on characterizing the association between previous fluconazole use and fluconazole resistance. Am J Infect Control 2010;38:456e60. [34] Wang JS, Li SY, Yang YL, Chou HH, Lo HJ. Association between fluconazole susceptibility and genetic relatedness among Candida tropicalis isolates in Taiwan. J Med Microbiol 2007; 56:650e3. [35] van Asbeck EC, Huang YC, Markham AN, Clemons KV, Stevens DA. Candida parapsilosis fungemia in neonates: genotyping results suggest healthcare workers hands as source, and review of published studies. Mycopathologia 2007;164:287e93. [36] Li SY, Yang YL, Lin YH, Ko HC, Wang AH, Chen KW, et al. Two closely related fluconazole-resistant Candida tropicalis clones circulating in Taiwan from 1999 to Microb Drug Resist 2009;15:205e10. [37] Pfaller MA, Diekema DJ, Rex JH, Espinel-Ingroff A, Johnson EM, Andes D, et al. Correlation of MIC with outcome for Candida species tested against voriconazole: analysis and proposal for interpretive breakpoints. J Clin Microbiol 2006; 44:819e26. [38] Pfaller MA, Messer SA, Boyken L, Rice C, Tendolkar S, Hollis RJ, et al. Cross-resistance between fluconazole and ravuconazole and the use of fluconazole as a surrogate marker to predict susceptibility and resistance to ravuconazole among 12,796 clinical isolates of Candida spp. J Clin Microbiol 2004;42:3137e41. [39] Pfaller MA, Messer SA, Boyken L, Rice C, Tendolkar S, Hollis RJ, et al. Use of fluconazole as a surrogate marker to predict susceptibility and resistance to voriconazole among 13,338 clinical isolates of Candida spp. tested by clinical and laboratory standards institute-recommended broth microdilution methods. J Clin Microbiol 2007;45:70e5. [40] Pfaller MA, Messer SA, Boyken L, Tendolkar S, Hollis RJ, Diekema DJ. Selection of a surrogate agent (fluconazole or voriconazole) for initial susceptibility testing of posaconazole against Candida spp.: results from a global antifungal surveillance program. J Clin Microbiol 2008;46:551e9.