RESUMEN
The identification of microorganisms at the species level has always constituted a diagnostic challenge for clinical microbiology laboratories. The aim of the present study has been the evaluation in a real-time assay of the performance of Autobio in comparison with the Bruker mass spectrometry system for the identification of bacteria and yeasts. A total of 535 bacteria and yeast were tested in parallel with the two systems by direct smear or fast formic acid extraction for bacteria and yeasts, respectively. Discordant results were verified by 16S, ITS rRNA or specific gene sequencing. Beyond giving comparable results for bacteria with respect to the MBT smart system, Autof MS2600 mass spectrometer provided excellent accuracy for the identification of yeast species of clinical interest.
RESUMEN
The Fourier-transform infrared spectroscopy-based IR Biotyper is a straightforward typing tool for bacterial species, but its use with Candida species is limited. We applied IR Biotyper to Candida parapsilosis, a common cause of nosocomial bloodstream infection (BSI), which is aggravated by the intra-hospital spread of fluconazole-resistant isolates. Of 59 C. parapsilosis isolates studied, n = 56 (48 fluconazole-resistant and 8 fluconazole-susceptible) and n = 3 (2 fluconazole-resistant and 1 fluconazole-susceptible) isolates, respectively, had been recovered from BSI episodes in 2 spatially distant Italian hospitals. The latter isolates served as an outgroup. Of fluconazole-resistant isolates, n = 40 (including one outgroup) harbored the Y132F mutation alone and n = 10 (including one outgroup) harbored both Y132F and R398I mutations in the ERG11-encoded azole-target enzyme. Using a microsatellite typing method, which relies on the amplification of genomic short tandem repeats (STR), two major clusters were obtained based on the mutation(s) (Y132F or Y132F/R398I) present in the isolates. Regarding IR Biotyper, each isolate was analyzed in quintuplicate using an automatic (i.e., proposed by the manufacturer's software) or tentative (i.e., proposed by us) cutoff value. In the first case, four clusters were identified, with clusters I and II formed by Y132F or Y132F/R398I isolates, respectively. In the second case, six subclusters (derived by the split of clusters I and II) were identified. This allowed to separate the outgroup isolates from other isolates and to increase the IR Biotyper typeability. The agreement of IR Biotyper with STR ranged from 47% to 74%, depending on type of cutoff value used in the analysis. IMPORTANCE Establishing relatedness between clinical isolates of Candida parapsilosis is important for implementing rapid measures to control and prevent nosocomial transmission of this Candida species. We evaluated the FTIR-based IR Biotyper, a new typing method in the Candida field, using a collection of fluconazole-resistant C. parapsilosis isolates supposed to be genetically related due to the presence of the Y132F mutation. We showed that IR Biotyper was discriminatory but not as much as the STR method, which is still considered the method of choice. Further studies on larger series of C. parapsilosis isolates or closely related Candida species will be necessary to confirm and/or extend the results from this study.
RESUMEN
Candida auris and other Candida species (C. albicans, C. glabrata, C. parapsilosis, C. tropicalis, and C. krusei) are important causes of bloodstream infection. Early or prolonged treatment with antifungal agents is often required. The inhibitory effect of antifungal agents in the patients' bloodstream may compromise the sensitivity of blood culture (BC) to diagnose and/or monitor patients with candidemia. Using a clinical BC simulation model, we compared antimicrobial drug-neutralizing BC media in BacT/Alert FA PLUS (FAP) or Bactec Plus Aerobic/F (PAF) bottles with non-neutralizing BC media in Bactec Mycosis IC/F (MICF) bottles to allow Candida growth in the presence of 100%, 50%, or 25% peak serum level (PSL) antifungal concentrations. In total, 117 organism/antifungal combinations were studied, and Candida growth was detected after incubating bottles into BacT/Alert VIRTUO or Bactec FX BC systems. Compared to control (without antifungal) bottles, both FAP and PAF bottles with 100% PSL antifungal concentrations allowed 100% recovery for C. auris, C. glabrata, and C. parapsilosis, whereas recovery was below 100% for C. albicans, C. krusei, and C. tropicalis. MICF bottles were less efficient at 100%, 50%, or 25% PSL antifungal concentrations, for all Candida species, except for C. auris. While azoles and amphotericin B did not hinder Candida growth in FAP or PAF bottles, echinocandins allowed C. auris, C. glabrata, and C. parapsilosis to grow in FAP, PAF, or MICF bottles. Overall, the maximum time to detection was 4.6 days. Taken together, our findings emphasize the reliability of BCs in patients undergoing antifungal treatment for candidemia. IMPORTANCE While echinocandins remain the preferred antifungal therapy for candidemia, bloodstream infections caused by C. auris, C. glabrata, or, at a lesser extent, C. parapsilosis may be difficult to treat with these antifungal agents. This is in view of the high propensity of the above-mentioned species to develop antifungal resistance or tolerance during treatment. Azoles and amphotericin B are possible alternatives. Thus, optimizing the recovery of Candida from BCs is important to exclude the likelihood of negative BCs for Candida species, owing to the inhibitory effect of antifungal agents present in the blood sample with which BCs are inoculated. Consistently, our results about the recovery of medically important Candida species (including C. auris) from simulated BCs in BacT/Alert FAP, Bactec PAF, or Bactec MICF bottles containing clinically relevant antifungal concentrations add support to this research topic, as well as to the use of BCs for monitoring the clinical and therapeutic course of candidemia.
