Evaluation of 11 DNA Automated Extraction Protocols for the Detection of the 5 Mains Candida Species from Artificially Spiked Blood

The molecular detection of Candida plays an important role in the diagnosis of candidaemia, a major cause of morbidity and mortality. The sensitivity of this diagnosis is partly related to the efficiency of yeast DNA extraction. In this monocentric study, we investigated the suitability of 11 recent automated procedures for the extraction of low and high amounts of Candida DNA from spiked blood. The efficacy of the DNA extraction procedures to detect Candida spp. in blood samples ranged from 31.4% to 80.6%. The NucliSENS(TM) easyMAG(TM) procedure was the most efficient, for each species and each inoculum. It significantly outperformed the other procedures at the lower Candida inocula mimicking the clinical setting. This study highlighted a heterogeneity in DNA extraction efficacy between the five main Candida species (Candida albicans, Candida glabrata, Candida parapsilosis, Candida tropicalis and Candida krusei). Up to five automated procedures were appropriate for C. krusei DNA extraction, whereas only one method yielded an appropriate detection of low amount of C. tropicalis. In the era of the syndromic approach to bloodstream infection diagnosis, this evaluation of 11 automated DNA extraction methods for the PCR diagnosis of candidaemia, puts the choice of an appropriate method in routine diagnosis within the reach of laboratories.