Misidentification of Candida Species by Clinical Laboratories— Not a Good Thing!

Abstracts & Commentary

Synopsis: Almost 1 in 20 Candida species isolated were misidentified by clinical laboratories; these errors have potential for engendering adverse clinical outcomes. C glabrata was the species most frequently misidentified.

Sources: Coignard C, et al. Resolution of discrepant results for Candida species identification by using DNA probes. J Clin Microbiol. 2004;42:858-861; Hajjeh RA, et al. Incidence of bloodstream infections due to Candida species and in vitro susceptibilities of isolates from 1998 to 2000 in a population-based active surveillance program. J Clin Microbiol. 2004;42:1519-1527.

Coignard and colleagues at the CDC examined 935 Candida bloodstream isolates from 51 institutions in Connecticut and the city and county of Baltimore. They used a DNA probe-based species identification system (PCR-EIA) in order to resolve discrepancies in phenotypic identification by more classical methods between the referring institutions and CDC.

Twenty-three (54%) of the referring institutions used germ-tube formation tests to identify Candida albicans, and 80% used one or another carbon assimilation biochemical panels to identify non-albicans species. The CDC used the CHROMagar, API 20C AUX (also used by 39% of referring sites) or RaPID Yeast Plus system, and microscopic morphology on cornmeal-Tween 80 plates.

CDC’s phenotypic identification of all but one isolate was confirmed by PCR-EIA, but identification discrepancies between CDC and the referring institutions were observed with 43 (4.6%) of the isolates (see Table, below). C albicans comprised 45% of isolates, C glabrata 24%, and C parapsilosis 13%. C glabrata was the organism that had most frequently been misidentified, accounting for 37% of discrepancies, followed by C parapsilosis (35%). Seven percent of C glabrata and 12% of C parapsilosis were misidentified. Of the 16 C glabrata isolated and for which a discrepancy was observed, 7 (44%) were misidentified as C albicans. Six C parapsilosis were misidentified as C albicans, as was one C lusitaniae.

Hajjeh and colleagues at the CDC, Yale, and Johns Hopkins examined the susceptibility of this group of isolates to antifungal agents by NCCLS broth microdilution, except for amphotericin B, for which the E test was used. High-level resistance (MIC > 64 mg/mL) to fluconazole was detected in only 1.2% of C albicans and in 5.9% of non-albicans isolates. While all C parapsilosis isolates were susceptible, 7% of C glabrata and 6% of C tropicalis were resistant. Only 35% of C krusei isolates were considered resistant. Except for C albicans, the frequencies of itraconazole resistance were somewhat higher. The amphotericin B MIC was > 1 mg/mL for 1.7% of all isolates and was > 1 mg/mL for only 0.4%. The amphotericin B MIC90 for the 15 C lusitaniae isolates was 0.25 mg/mL, identical to that for C glabrata, while the MIC90 for C krusei was 2.0 mg/mL. Flucytosine performed well, with only 2.4% of all isolates resistant.

Comment by Stan Deresinski, MD, FACP

While the overall identification error rate of 4.6% does not seem terrible, severe consequences could potentially result from some of these misidentifications. Because antifungal susceptibility testing is not readily available to many clinicians (susceptibility testing was performed at the referring institution on only 10% of isolates) and, even when it is, the results are often not available for a week or more, there is a frequent reliance upon species identification in predicting likely antifungal susceptibility patterns and antifungal therapy. Thus, an identification of an isolate as C albicans usually implies to the clinician a relatively high likelihood of susceptibility to azole antifungal agents. Initial identification by the laboratory of "Candida, non-albicans" in contrast, raises increased concern about azole resistance.

Species identification of some non-albicans isolates may further increase that concern. For example, C krusei is considered invariably resistant to high concentrations of fluconazole (although this was not found with the isolates studied here). C glabrata, which is being isolated with increasing frequency from the bloodstream while the incidence of C albicans bloodstream infection is decreasing,1 and which comprised 24% of the bloodstream isolates in this study, is also more frequently resistant to fluconazole. While 7% of C glabrata in this study had a fluconazole MIC > 64 mg/mL, a recent nationwide survey of 559 C glabrata isolates reported a 9% incidence of resistance; 31% were "susceptible-dose dependent" with MICs 16-32 mg/mL, a category not reported here by the CDC.2 In addition, there was a striking geographic variation in resistance rates ranging from none in New England (a result not reflected in the Connecticut data here) to 15% in the Mid-Atlantic states.2 C glabrata can be problematic in other ways. It evidences delayed growth in blood culture, a feature that could delay institution of antifungal therapy.3 Similar delayed growth in urine culture can lead to missing its presence altogether in consequence of the early discarding of culture plates. Furthermore, misidentification of C glabrata as C albicans in blood could lead to inappropriate antifungal therapy with potentially severe consequences.

The changing epidemiology of invasive Candida infections, in particular the increasing proportion due to non-albicans strains, has important consequences for patient management. This change puts increasing demands upon our clinical laboratories, clinicians, and pharmacy budgets. Clinicians must have an increased level of suspicion of the presence of invasive candidiasis, particularly in cases without positive blood cultures or in critically ill patients for whom the delay in institution of antifungal therapy until blood cultures turn positive could prove lethal. Laboratories must ensure the rapid isolation and accurate identification of infecting pathogens, as well as the availability of timely availability of antifungal susceptibility data. Pharmacy budgets are in trouble.


1. Trick WE, et al. Secular trend of hospital-acquired candidemia among intensive care unit patients in the United States during 1989-1999. Clin Infect Dis. 2002;35:627-630.

2. Pfaller MA, et al. Variation in susceptibility of bloodstream isolates of Candida glabrata to fluconazole according to patient age and geographic location. J Clin Microbiol. 2003;41:2176-2179.

3. Horvath LL, et al. Detection of simulated candidemia by the BACTEC 9240 system with plus aerobic/F and anaerobic/F blood culture bottles. J Clin Microbiol. 2003;41:4714-4717.

Stan Deresinski, MD, FACP Clinical Professor of Medicine, Stanford; Associate Chief of Infectious Diseases, Santa Clara Valley Medical Center, is Editor of Infectious Disease Alert.