Rapid and cost-effective identification of Candida species using multipoint inoculation of CHROMagar Candida media, cycloheximide sensitivity and carbohydrate assimilation tests

Abstract

<h2>Summary</h2><h3>Aims</h3> Increasing numbers of nosocomial infections with <i>Candida</i> species are caused by species other than <i>Candida albicans.</i> Definitive identification of <i>Candida</i> spp. usually utilises one commercial biochemical identification strip per isolate. As an alternative, CHROMagar Candida (CHROMagar, France) can be used to differentiate some species but lacks accuracy for some species, and is expensive when used as a primary isolation medium. The aim of this study was to evaluate a multipoint inoculation method for batch identification of <i>Candida</i> isolates, by generating a unique profile using CHROMagar Candida, a cycloheximide-containing medium and media containing seven in-house prepared carbohydrates. <h3>Methods</h3> Initially, 53 prospectively obtained clinical isolates, 52 stored clinical isolates and 15 reference laboratory-derived strains were tested using the multipoint inoculation method and compared with results obtained using the ID32C (BioMerieux, France) plus a germ tube test and yeast morphology agar. Nine <i>Candida</i> spp. were represented. Subsequently, 400 <i>Candida</i> isolates obtained from clinical specimens were prospectively tested using this method. <h3>Results</h3> All <i>Candida glabrata</i> (37), <i>Candida parapsilosis</i> (27), <i>Candida albicans</i> (26), <i>Candida tropicalis</i> (11), <i>Candida lusitaniae</i> (three), <i>Candida kefyr</i> (three) and <i>Candida famata</i> (two) were correctly identified. Of 26 <i>C. albicans</i> isolates, 25 gave a green colour on CHROMagar Candida and one gave a less characteristic white colour. One of five <i>Candida guilliermondii</i> failed to assimilate cellobiose. All six <i>Candida krusei</i> isolates consistently produced a unique unexpected profile which was then incorporated into the database and the method was integrated into the routine laboratory. Of 400 consecutive prospectively examined isolates, 390 (97.5%) were identifiable using this system; 53.7% were due to Candida species other than <i>C albicans.</i> Seven isolates (1.8%) were not identifiable using this system as they did not produce any of the expected profiles. <h3>Conclusions</h3> Utilising the described method and batch testing allowed considerable cost and labour savings for the laboratory identification of <i>Candida</i> species and determination of local <i>Candida</i> epidemiology. The estimated material and labour costs for a run of 30 yeasts in a batch is AUD$0.97. By comparison, the cost of a commercial ID32C is AUD$8.52 per isolate.