Surveillance of the prevalence, antibiotic susceptibility, and genotypic characterization of invasive candidiasis in a teaching hospital in China between 2006 to 2011

Background Invasive candidiasis is an important nosocomial infection associated with high mortality among immunosuppressive or critically ill patients. We described the incidence of invasive candidiasis in our hospital over 6 years and showed the antifungal susceptibility and genotypes of the isolated yeast. Method The yeast species were isolated on CHROMagar Candida medium and identified using an yeast identification card, followed by analysis of the D1/D2 domain of 26S rDNA. The susceptibilities of the isolates to flucytosine, amphotericin B, fluconazole, itraconazole, and voriconazole were tested using the ATB FUNGUS 3 system, and that to caspofungin was tested using E-test strips. C. albicans was genotyped using single-strand conformation polymorphism of CAI (Candida albicans I) microsatellite DNA combined with GeneScan data. Results From January 2006 to December 2011, a total of 259 isolates of invasive Candida spp. were obtained from 253 patients, among them 6 patients had multiple positive samples. Ninety-one stains were from blood and 168 from sterile fluids, accounting for 6.07% of all pathogens isolated in our hospital. Most of these strains were C. albicans (41.29% in blood/59.06% in sterile body fluids), followed by C. tropicalis (18.06%/25.72%), C. parapsilosis (17.42%/5.43%), C. glabrata (11.61%/3.99%) and other Candida spp. (11.61%/5.80%). Most Candida spp. were isolated from the ICU. The new species-specific CLSI candida MIC breakpoints were applied to these date. Resistance to fluconazole occurred in 6.6% of C. albicans isolates, 10.6% of C. tropicalis isolates and 15.0% of C. glabrata isolates. For the 136 C. albicans isolates, 54 CAI patterns were recognized. The C. albicans strains from blood or sterile body fluids showed no predominant CAI genotypes. C. albicans isolates from different samples from the same patient had the same genotype. Conclusions Invasive candidiasis has been commonly encountered in our hospital in the past 6 years, with increasing frequency of non-C. albicans. Resistance to fluconazole was highly predictive of resistance to voriconazole. CAI SSCP genotyping showed that all C. albicans strains were polymorphic. Invasive candidiasis were commonly endogenous infection.


Background
In the early 1980s, fungi have emerged as major causes of human disease, especially in immunocompromised and hospitalized patients [1]. Candida spp. are the fourth most common pathogens in nosocomial bloodstream infections (BSIs), and they account for 12% of all hospital-acquired BSIs in the United States [2]. The mortality of candidemia in adults and neonatal children was estimated to be 15%-25% and 10%-15% in the United States, respectively [3]. Further, studies have shown that Candida infection can extend the duration of hospital stay [4], increase the cost of medical care [5], and lead to a high mortality (30%-40%) [6]. In recent years, non-C. albicans spp. have been detected more commonly in candidemia [7,8]. Both C. albicans and non-C. albicans cause invasive candidiasis, and the antifungal resistance of invasive Candida aggravates the situation. Some reports on Candida resistance to antifungals showed that resistance could make therapy more difficult [9,10]. The species and antifungal resistance of Candida differ among geographies, and it is necessary to survey the prevalence and drug resistance of Candida spp. in different locations in order to ensure effective antifungal therapy against invasive Candida.
Few studies on the long-term monitoring of the incidence and antifungal susceptibilities of invasive Candida spp. have been reported from China. This study aims to describe the trend of invasive candidiasis in a teaching hospital during the last 6 years.

Study population and data collection
The study was conducted in Beijing Chaoyang Hospital, a teaching facility of the Capital Medical University in Beijing, China. According to the diagnostic criteria, all patients with invasive candidiasis were enrolled from different wards from 2006 to 2011. The following information was collected including age, sex, underlying diseases, et al. This investigation had the approval of the Ethics Committee of Beijing Chaoyang Hospital and informed consent was obtained from all study subjects or their next of kin.

Candida strain collection
A total of 259 isolates of invasive Candida spp. were obtained from 253 patients, among them 6 patients had multiple positive samples. Invasive Candida strains were isolated from blood (91) and sterile samples (168) (including ascitic fluid, bile,central venous catheter, pleural fluid); the total of 259 samples accounted for 6.07% of all pathogens (5.75% of 1583 from blood and 6.27% of 2684 pathogens from sterile fluids) isolated in Beijing Chao-yang Hospital during the study period. Patients come from 13 provinces in china, mainly in northern china (including Beijing, Shanxi, Inner Mongolia, Heilongjiang et al.) Most of the Candida spp. were isolated from patients in the intensive care unit (ICUs). The Candida strains were primarily cultured on CHROMAgar Candida medium (JinZhang, Tianjing China) and identified using an yeast identification card (BioMérieux, France) or the API 20C AUX system (BioMérieux, France), followed by sequence analysis of the D1/D2 domain in the 26S rDNA [11]. Quality control isolates included C. albicans ATCC90028, C. parapsilosis ATCC22019, and C. krusei ATCC6258.

Antifungal drug susceptibility test
A total of 240 Candida strains were tested for antifungal susceptibility. The susceptibility of these strains to 6 antifungal drugs was tested in vitro. Flucytosine, amphotericin B, fluconazole, voriconazole, and itraconazole were tested using the ATB FUNGUS 3 system (1000723260; BioMérieux, France), and caspofungin, using the E-test (BJ3625 AB Biodisk; Solna, Sweden). For fluconazole, voriconazole, and itraconazole, the minimal inhibitory concentration (MIC) was determined by determining the concentration at which a prominent reduction in the yeast cell count was observed after 24 h of treatment. For the caspofungin, the MIC was determined by detection of a significant reduction in growth after 24 h of incubation. The MIC for amphotericin B and flucytosine were defined as the lowest concentration at which no visible growth was detected. Recently approved CLSI 24 h resistance breakpoints for fluconazole,voriconazole and echinocandins were used [12]: isolates of C. albicans, C. tropicalis and C. parapsilosis with an MIC ≥ 8 μg/ml and isolates of C. glabrata with an MIC ≥ 64 μg/ml were considered resistant to fluconazole; isolates of C. albicans, C. tropicalis and C. parapsilosis with an MIC ≥ 1 μg/ml were considered resistant to voriconazole; isolates of C. albicans and C. tropicalis with an MIC ≥ 1 μg/ml were considered resistant to caspofungin, isolates of C. parapsilosis with an MIC ≥ 8 μg/ml were considered resistant to caspofungin, isolates of C. glabrata with an MIC ≥ 0.5 μg/ml were considered resistant to caspofungin. Data were reported as the MIC ranges; the MIC 50 and MIC 90 values; and the number of resistant (R) isolates.

DNA extraction and PCR
DNA extraction and PCR amplification of CAI microsatellite DNA were performed according to the methods of [13]. The microsatellite locus CAI was amplified by PCR (forward primer, 5′-ATG CCA TTG AGT GGA ATT GG-3′; reverse, 5′-AGT GGC TTG TGT TGG GTT TT-3′). For GeneScan analysis, the forward primer was fluorescently labeled with 6-carboxyfluorescein at the 5′ end. PCR amplification was performed in a thermocycler (ICycler; Bio-Rad, Hercules, CA) with the following program: initial denaturing step at 95°C for 4 min; 33 cycles of denaturation at 95°C for 30 s, annealing at 60°C for 30 s, and extension at 72°C for 1 min; and a final extension step at 72°C for 7 min.

Genotype analysis
Genotype analysis of the amplified CAI fragments was carried out using the methods described by [14]. The sizes of the PCR products were automatically determined using the ABI 370 genetic analyzer (Applied Biosystem, Foster City, CA). Because of the diploid nature of C. albicans, the CAI genotypes of the strains were designated by the number of trinucleotide repeat units in both alleles of the microsatellite locus [14].

Statistical analysis
Drug susceptibility testing was analyzed using WHONET 5.6. Data were analyzed with SPSS 17.0 (SPSS Inc., Chicago, IL). Categorical variables were compared using the chi-square or Fisher's exact test. A P value of 0.05 was considered significant.
The top 3 wards in which candidemia occurred were the ICUs (43.23%) and the hematology (16.13%) and the urology (10.32%). The wards in which Candida spp. were most frequently isolated from sterile fluids were the ICUs (41.91%) and the interventional radiology (18.15%) and the hepatobiliary (10.56%). Thus, our data showed that most patients with nosocomial candidasis were from the ICUs.

Susceptibility of antifungal drugs
The susceptibilities of the 240 yeast strains to the 6 antifungal drugs are shown in Table 1. Most Candida strains were sensitive to all the antifungal drugs, although C. glabrata and C. tropicalis showed decreased sensitivity to triazole drugs (less than 90%). The percentages of isolates in the R categories were 7.5% for fluconazole and 5.4% for voriconazole, respectively. Caspofungin and amphotericin B exhibited excellent antifungal activity against all Candida species. The MIC of fluconazole was ≥8 μg/ml for C. albicans isolates (6.6%) and that of voriconazole was ≥1 μg/ml (2.9%). Further, the MIC of fluconazole was ≥8 μg/ml for C. tropicalis isolates (10.6%) and that of voriconazole was ≥1 μg/ml (10.6%). The MIC of fluconazole was ≥64 μg/ml for C. glabrata isolates (15.0%). Three of 136 C. albicans isolates (approximately 2%) were resistant to both FLC (MIC ≥ 8 μg/ml) and VRC (MIC ≥ 1 μg/ml). Six of 59 C. tropicalis isolates (approximately 10%) had FLC MIC ≥ 8 μg/ml and VRC ≥ 1 μg/ml. Three of 20 C. glabrata isolates (approximately 15%) had FLC MIC ≥ 64 μg/ml and VRC ≥ 8 μg/ml. No isolate was resistant to all existing classes. Cross-resistance to triazole antifungal drugs was found in 2.9% C. albicans and 10.6% C. tropicalis isolates. Resistance to fluconazole was highly predictive of voriconazole resistance. No statistically significant differences in azole resistance (P > 0.05) were observed for C. albicans, C. tropicalis, and C. glabrata isolated from either blood or sterile fluids.

Discussion
This study described the distribution, in vitro antifungal susceptibilities, and genotyping results of invasive Candida spp. isolated between 2006 and 2011 from Beijing Chaoyang hospital in China.
The results showed that invasive candidiasis has frequently occurred in our hospital in the past 6 years. The average isolation rate of Candida spp. was 6.07% among all pathogens isolated from blood and sterile fluid samples. In most cases, the invasive candidiasis occurred in the ICUs, indicating that more attention must be paid to ICU patients in order to prevent them from contracting this disease. Although C. albicans was the most predominant species, the frequency at which non-albicans Candida were detected increased obviously over the study period. In our hospital, in more than half the recent cases of candidemia, the causative organisms were non-albicans Candida strains, C. tropicalis, C. parapsilosis, and C. glabrata were the most common among the non-albicans Candida spp. detected, in accordance with reports by Pereira GH, et al. [16,17]. In the case of the blood samples, C. albicans no longer comprised the majority of isolated but remained the most frequently isolated species (41.29%), followed by C. tropicalis (18.06%), C. parapsilosis    (17.42%), and C. glabrata (11.61%). In the case of the sterile samples, C. albicans was the most commonly isolated organism (59.06%), followed by C. tropicalis (25.72%), C. parapsilosis (5.43%), and C. glabrata (3.99%). It is reported that C. glabrata is No.4 among the most isolated species, which are similar to our findings, but different from those reported from Europe and the US [1,18,19]. C. abicans has traditionally been the leading cause of candidemia worldwide, with the proportion of C.abicans infections decreasing and proportion of non-C.abicans infections increasing, especially C. tropicalis, C. parapsilosis and C. glabrata. The trend has also been observed in this study, the proportion of C. abicans candidemia cases decreased from 54% to 41%. At the same time, the proportion of C. parapsilosis candidemia, C. glabrata candidemia and C. tropicalis candidemia cases rose from 12% to 17%, 7% to 12% and 15% to 18%, respectively. In the ARTEMIS DISK global antifungal surveillance project undertaken between 1997 and 2007, 90% of infections were found to be caused by 4 Candida species, i.e., C. albicans, C. glabrata, C. parapsilosis, and C. tropicalis [20]. This distribution is similar to our study, C. abicans, C.tropicalis,C. parapsilosis and C. glabrata still comprise 92.67% of all isolates, that has not substantially changed in 20 years in the US [21]. During the last decade, the frequency of detection of the latter 3 species has been increasing in many countries, including China [22][23][24]. The widespread use of fluconazole may change the epidemiology of infections by nonalbicans Candida, especially azole-resistant Candida spp. [25,26]. The shift to non-albicans Candida species is also significant because of the newest class of antifungal agents, so the caspofungin, are not currently recommended as primary therapy against C. parapsilosis, the third most common species.
Few studies have conducted long-term research on invasive Candida in China. The antifungal resistance rate in our study was higher than that reported by Guiyang in 2009 [27] and Hong Kong in 2011 [28]. However, none of the isolates in this study were resistant to caspofungin and amphotericin B. More than 90% of the C. albicans isolates were susceptible to all antifungal drugs between 2006 and 2011. We found that fluconazole resistance in C. albicans increased from 2.2% to 6.6% and in C.tropicalis, increased from 5.1% to 10.6%. There was also,a trend toward increased resistance to azoles in an Italian study [29]. We also found crossresistance in our study, in agreement with previous reports from other countries [9,30]. A surveillance from 41 institutions participated study had shown that resistance to fluconazole was highly predictive of resistance to voriconazole [22]. In our study, all of the C. tropicalis isolates resistant to FLC (MIC ≥ 8 μg/ml) were also resistant to VRC (MIC ≥ 1 μg/ml). The same result in C. glabrata isolates which had FLC MIC ≥ 64 μg/ml and VRC ≥ 8 μg/ml was also observed. Therefore, we inferred that resistance to fluconazole highly tend to predictive of voriconazole resistance for C. glabrata and C. tropicalis.
Fifty-four CAI patterns were recognized from the 136 C. albicans isolates. CAI SSCP genotyping showed that the C. albicans strains were polymorphic. C. albicans isolates in different samples from the same patient had the same genotype. Compared to Li's study [13], we found more patterns. These results implied that the genotypic distribution of C. albicans strains may be various in different samples. Isolates in different samples from the same patient showed exactly the same CAI SSCP pattern (e.g. isolates from patients CY61, CY65, CY66, CY69, CY141, CY119, CY123, CY68 and CY154 in Table 2), indicating that the infection was mainly endogenous [14,31].

Conclusions
In summary, this study showed the epidemiology and antifungal resistance of invasive Candida spp. isolated from Beijing Chaoyang hospital between 2006 and 2011. C. albicans was the most frequently isolated species, compared to the non-albicans Candida spp. Although the overall rates of resistance to antifungal agents remain low, species-specific raise some concerns, especially C. tropicalis and C. glabrata were the most frequent detected among organisms resistant to fluconazole. Currently, Caspofungin and amphotericin B are highly effective in vitro against all Candida species. Most of the invasive candidiasis cases occurred in the ICU, so careful attention should be paid to ICU patients to prevent them from contracting this disease. Genotyping of CAI SSCP showed that the C. albicans strains were polymorphic from blood and sterile samples and isolates from different samples of the same patient had the same genotype.