TABLE 5

Clinical studies evaluating Candida PCR a

Study (reference no.)Study designPatient populationType of PCRType of specimen testedPrimer targetMethod utilized to determine accuracySensitivity (%)Specificity (%)Comments
Ahmad et al., 2002 (185)Retrospective28 culture-proven or suspected Candida-positive patients, 10 superficially colonized patients, 12 healthy controlsSeminested PCRSerumITS2Comparison to blood culture results100100The approach was able to identify Candida species in 9 culture-negative patients with suspected IC. All Candida-colonized patients had negative PCR results.
White et al., 2003 (186)Prospective113 patients at risk for ICReal-time PCR and nested PCRWhole blood18S rDNANANANAOnly 3 of the 113 patients had blood culture-positive results, 2 of whom also had positive PCR results. PCR was positive in another 25 patients suspected to have IC.
Tirodker et al., 2003 (187)Prospective70 pediatric and neonatal ICU patients with sepsisTraditional PCR with gel electrophoresisWhole blood18S rDNAComparison to blood culture results10077.2Seven of 13 culture-negative and PCR-positive patients had other evidence of IFI.
Maaroufi et al., 2003 (118)Retrospective61 hemato-oncology patients with proven or suspected ICReal-time PCR with hydrolysis probesWhole blood5.8S and 28S rDNAComparison to blood culture results10097
Maaroufi et al., 2004 (188)Retrospective39 patients with clinically proven or suspected Candida infection and 15 controlsReal-time PCR with fluorescent probesSerum5.8S and 28S rDNAComparison to blood culture results10097
Ahmad et al., 2004 (189)Retrospective26 patients (6 proven cases, 10 suspected cases, 10 healthy controls)Seminested PCR-ELISASerum5.8S and 28S rDNA and ITS2Comparison to blood culture results10080PCR was positive in 4 of 10 patients with suspected IC and in none of the healthy controls.
White et al., 2005 (190)Retrospective105 patients at high risk for IFIReal-time PCR with fluorescent probeWhole blood and serum18S rDNAComparison to EORTC/MSG criteria95 for proven and probable cases97Possible IC cases were not included in the sensitivity and specificity determinations.
Moreira-Oliveira et al., 2005 (191)Prospective225 patients with hematologic malignancies and at risk for ICTraditional PCR followed by sequencingWhole blood5.8S rDNAComparison to blood culture results72.191.2
Alam et al., 2007 (192)Retrospective27 patients with culture-proven Candida infection, 39 patients with suspected candidemia, 10 colonized patients, 16 controlsSeminested PCRSerumITS2Comparison to EORTC/MSG criteria92.5100Probable IC cases were excluded from analysis (53% of them were PCR positive).
McMullan et al., 2008 (193)Prospective157 nonneutropenic patients in the ICUReal-time PCR with hydrolysis probesSerum18S and 5.8S rDNA, ITS1, ITS2Comparison to EORTC/MSG criteria, modified for nonneutropenic patients82100Probable IC patients were excluded from the analysis. One of the 11 patients with proven IC was diagnosed with Candida famata infection, which was not possible to detect with the primers used.
Dunyach et al., 2008 (194)Prospective23 ICU patients with cancer and 10 healthy controlsReal-time PCR with SYBR green fluorescenceSerumITS1 to ITS4, L18Comparison to blood culture results92 for L18 PCR, 76.9 for ITS PCR66 for L18 PCR, 100 for ITS PCRPatients with probable IC were excluded from the analysis. Among the 10 patients with probable IC, 3 were positive with L18 PCR and 5 with ITS PCR.
Metwally et al., 2008 (195)Retrospective104 patients included in a previous prospective study, from whom whole-blood specimens were obtainedReal-time PCR with hydrolysis probesWhole blood and serum18S and 5.8S rDNA, ITS1, ITS2Comparison to EORTC/MSG criteria, modified for nonneutropenic patients100 for serum, 70 for whole blood100
Badiee et al., 2009 (196)Prospective194 patients with hematologic malignanciesPCR-ELISAWhole blood18S rDNAComparison to EORTC/MSG criteria100 (2 of 2 proven IC cases)95Patients with probable IC and patients with fever of unknown origin were excluded from the analysis. PCR remained positive until death when treatment failed.
Khlif et al., 2009 (197)Prospective110 patients at risk for ICReal-time and nested PCR assaysBlood cultures18S and 5.8S rDNA, ITS1, ITS2Comparison to blood cultures81 for real-time PCR, 86 for nested PCR96 for real-time PCR, 54 for nested PCR
Wellinghausen et al., 2009 (198)Prospective284 patients at risk for ICReal-time PCRWhole blood18S rDNAComparison to blood cultures87.593PCR also detected Candida DNA in 8 blood culture-negative patients with Candida isolated from culture-sterile sites.
Badiee et al., 2010 (199)Prospective35 patients with bone marrow transplantReal-time PCR with hydrolysis probesWhole blood18S rDNAComparison to EORTC/MSG criteria100 (probable IC cases were considered truly positive)88.9 (increased to 100 when only patients with at least two positive PCR results were considered PCR positive)
Lau et al., 2010 (200)Retrospective109 patients with or at risk for candidemiaMultiplex tandem PCRWhole bloodITS1, ITS2, elongation factor 1a, β-tubulinComparison to blood cultures7597Results were accelerated by an average of 2.2 days compared to culture. Serum and plasma PCRs were more sensitive with the few serum samples that were tested.
Schell et al., 2012 (201)Retrospective16 patients with culture-proven Candida infectionReal-time and microfluidic PCRWhole bloodITS1, ITS2Comparison to blood cultures68.7 for real-time PCR, 56.2 for microfluidic PCRNA due to study design
Trovato et al., 2012 (202)Retrospective86 neonatal ICU patients with suspected bloodstream infectionsPCR followed by ethidium bromide stainingBlood cultures18S rDNA, ITS1, 28S rDNAComparison to EORTC/MSG criteria87.5 for proven and probable IC98.6 for no ICFor comparison, the sensitivity and specificity of the blood culture results were 50% and 100%, respectively.
Nguyen et al., 2012 (4)Retrospective55 patients with IC and 73 hospitalized controlsReal-time PCR with fluorescently labeled probesWhole blood, plasma, serumITS1, ITS2IC was defined as recovery of Candida from blood or a sterile site; controls were defined as those having no clinical or microbiological evidence of IC8070“Positive PCR” was defined as one positive plasma or serum PCR result. In a preliminary run, whole-blood PCR was found to have a significantly lower sensitivity than that of plasma or serum PCR. The β-d-glucan sensitivity and specificity were 56% and 73%, respectively.

  • ↵a EORTC/MSG, European Organization for Research and Treatment of Cancer/Mycoses Study Group; IC, invasive candidiasis; ICU, intensive care unit; IFI, invasive fungal infection; ITS, internal transcriber spacer; NA, not applicable.