A Rapid and Automated Sample-to-Result Candida auris Real-Time PCR Assay for High-Throughput Testing of Surveillance Samples with the BD Max Open System

The multidrug-resistant yeast pathogen Candida auris continues to cause outbreaks and clusters of clinical cases worldwide. Previously, we developed a real-time PCR assay for the detection of C. auris from surveillance samples (L. Leach, Y. Zhu, and S. Chaturvedi, J Clin Microbiol 56:e01223-17, 2018, https://doi.org/10.1128/JCM.01223-17).

ABSTRACT

The multidrug-resistant yeast pathogen Candida auris continues to cause outbreaks and clusters of clinical cases worldwide. Previously, we developed a real-time PCR assay for the detection of C. auris from surveillance samples (L. Leach, Y. Zhu, and S. Chaturvedi, J Clin Microbiol 56:e01223-17, 2018, https://doi.org/10.1128/JCM.01223-17). The assay played a crucial role in the ongoing investigations of the C. auris outbreak in New York City. To ease the implementation of the assay in other laboratories, we developed an automated sample-to-result real-time C. auris PCR assay using the BD Max open system. We optimized sample extraction at three different temperatures and four incubation periods. Sensitivity was determined using eight pools of patient samples, and specificity was calculated using four clades of C. auris and closely and distantly related yeasts. Three independent extractions and testing of two patient sample pools in quadruplicate yielded assay precision. BD Max optimum assay conditions were as follows: DNA extraction at 75°C for 20 min and the use of PerfeCTa multiplex quantitative PCR (qPCR) ToughMix. The limit of detection (LOD) of the assay was one C. auris CFU/PCR. We detected all four clades of C. auris without cross-reactivity to other yeasts. Of the 110 patient surveillance samples tested, 50 were positive for C. auris using the BD Max system with 96% clinical sensitivity and 94% accuracy compared to the results of the manual assay. The BD Max assay allows high-throughput C. auris screening of 180 surveillance samples in a 12-h workday.

INTRODUCTION

Candida auris, an emerging multidrug-resistant yeast pathogen, continues to cause outbreaks and clusters of clinical cases worldwide (1, 2). There are ongoing efforts to devise better diagnostic approaches for the rapid detection of C. auris in clinical and surveillance samples (3–5). Previously, we developed and validated a manual real-time PCR assay for the direct detection of C. auris from surveillance samples at the New York State Department of Health (NYSDOH) Mycology Laboratory (3). The laboratory-developed test (LDT) enabled NYSDOH laboratory scientists and epidemiologists to carry out unprecedented surveillance and testing in hospitals and health care facilities in New York City. To date, over 13,000 clinical samples from 151 health care facilities and over 1,000 C. auris isolates were processed (S. Chaturvedi, unpublished data). The Candida auris LDT, standard operating procedures, and validation results were shared extensively with hospital, commercial, and public health laboratories. However, the C. auris LDT is not amenable to automation and high-throughput screening, and adoption of the LDT has progressed slowly while the affected facilities and sample numbers continue to grow. Therefore, we developed and validated a real-time PCR assay for C. auris using the BD Max system, a fully integrated and automated platform amenable to direct detection of other fungal pathogens (6, 7).

MATERIALS AND METHODS

As we aimed to migrate the manual C. auris real-time PCR assay to BD Max, the primers, probe, and PerfeCTa multiplex quantitative PCR (qPCR) ToughMix (Quanta BioSciences) were evaluated with the BD Max DNA MMK SPC master mix (3). We optimized the BD Max ExK DNA-1 (plasma/serum/urine) DNA extraction kit on the BD Max using three different temperatures and four incubation periods. We carried out the entire sample-to-result procedure in the BD Max PCR cartridges. BD Max assay sensitivity was determined using eight pools of patient surveillance samples (axilla, groin, axilla-groin, and nares swabs) in parallel testing with the manual real-time PCR assay. Pooled samples were used to ensure sufficient volume to test identical samples in triplicate using the manual and BD Max assays. High, medium, and low threshold cycle (C_T) pools were created by combining equal volumes from individual swab samples with similar C_T values based on the manual assay. All pooled samples were then rerun with the manual assay and the BD Max assay to be consistent with the results. We determined assay specificity by the analysis of four clades of C. auris and closely and distantly related yeasts at high and low concentrations. The primers and probe set were previously screened against a more extensive panel comprising closely and distantly related fungi, bacteria, viruses, and parasites (3). Assay precision was determined by testing samples in three independent extractions and testing two patient pools in quadruplicate. Following validation, we used the assay for the screening of 110 individual patient surveillance samples. GraphPad Prism 8 software (GraphPad Software, Inc., La Jolla, CA) was used for statistical analysis. We used Student’s t test for analysis of the means and considered a P value of <0.05 statistically significant. The manual C. auris real-time PCR assay was selected as the “gold standard” to assess the diagnostic performance of the BD Max assay for the surveillance samples. BD Max assay hands-on time estimation involved the sum of the time required to complete sample preparation, device loading and cleaning, and result review and reporting.

RESULTS

The initial experiments involved optimization of DNA extraction and amplification steps with the BD Max system. This included initial DNA extraction at three different temperatures, 70°C, 75°C, and 80°C; five different incubation times ranging from 10 to 30 min; and amplification using two master mixes, the PerfeCTa multiplex qPCR ToughMix (manual real-time PCR assay) and BD Max DNA MMK SPC master mix. The ToughMix appeared to be better than the BD Max DNA MMK SPC master mix, as it allowed more efficient amplification of C. auris DNA (see Fig. S1A in the supplemental material). Next, we fine-tuned DNA extraction time and temperature using the ToughMix. Our results revealed that the optimum sample DNA extraction condition was 75°C for 20 min on the BD Max (Fig. S1B). The selected extraction condition was further evaluated using two pooled surveillance samples (axilla, groin, axilla-groin, and nares) with high and medium C_T values in quadruplicate, and results showed excellent reproducibility (see Table S1 in the supplemental material). In expanded testing, the BD Max assay was highly sensitive with the limit of detection (LOD) of one C. auris CFU/PCR (Fig. 1). The assay detected all four known clades of C. auris with similar C_T values for the same concentration of cells (see Table S2 in the supplemental material). The assay was highly specific, as no cross-reactivity was seen against two closely related yeasts, Candida haemulonii and Candida duobushaemulonii, or 2 distantly related yeasts, Candida albicans and Candida glabrata, tested at high (1 × 10⁴ cells/PCR) and low (1 × 10¹ cells/PCR) concentrations (Table S2). The assay was highly reproducible, as it produced consistent C_T values on three different days of testing (see Table S3A in the supplemental material), and within the same day of testing (Table S3B). Upon completion of the test validation, we tested 110 individual patient surveillance samples; 50 were positive for C. auris by the BD Max system, resulting in 96% clinical sensitivity and 94% accuracy when compared to the results of the manual real-time PCR assay (Table 1). The C_T value distributions of 110 individual patient surveillance samples showed excellent correlation between the two methods (Fig. 2). We estimated the automated sample-to-result C. auris BD Max assay to allow the screening of 180 samples in a 12-h workday.

FIG 1.

BD Max Candida auris assay sensitivity. The pools of patient surveillance samples ranging in C. auris CFU from 5 × 10⁻¹ to 5 × 10⁵/PCR were run in duplicate on three different days. The assay was linear over 6 orders of magnitude, and the limit of detection was 1 CFU/PCR on the BD Max and manual assay.

TABLE 1.

Comparison of test performance of individual patient surveillance samples

BD Max real-time PCR result	Manual real-time PCR result		Accuracy (%)	Sensitivity (%)	Specificity (%)	PPVa (%)	NPVa (%)
	No. positive	No. negative
No. positive	45	5	94	96	92	92	97
No. negative	2	58

^aPPV, positive predictive value; NPV, negative predictive value.

FIG 2.

Side-by-side comparison between Candida auris BD Max real-time PCR and manual real-time PCR assays. C_T distribution of 110 individual patient surveillance samples showed that the BD Max real-time PCR assay had an excellent correlation for the detection of C. auris DNA with the manual real-time PCR assay for most of the samples (r = 0.9955).

DISCUSSION

We developed and validated the first automated sample-to-result real-time PCR assay for high-throughput testing of C. auris. The highlights of this study are as follows: easy migration of the manual real-time PCR assay to an automated platform; the consistent performance of primers, probe, and master mix from our manual assay; comparable accuracy, sensitivity, and specificity; and easy integration in the laboratory workflow. These findings also confirm the suitability of the BD Max platform for direct detection of fungal pathogens from clinical specimens as reported earlier for Pneumocystis jirovecii and Coccidioides immitis (6, 7). There was indication that the BD Max performed better than the manual assay for a few samples with high C_T values, possibly due to the elimination of many hands-on steps and the use of purified DNA. Although the samples tested in the present study belonged to C. auris South Asia clade I, our primers and probe have performed equally well with other well-known C. auris clades in the validation panel and the manual assay described earlier (3). Ahmad et al. partially modified our manual real-time PCR assay by carrying out DNA extraction from suspected C. auris surveillance samples on the MagNA Pure 96 automated extraction system (Roche Diagnostic System, Indianapolis, IN, USA); these authors reported a good correlation of the modified test findings with the results obtained by cell culture and matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF-MS) identification (4). The authors further claimed that partial automation of the real-time PCR assay allowed testing of approximately 200 samples per day. However, the trend in infectious disease diagnosis is toward full automation, multiplexing, and miniaturization of assay systems to facilitate rapid point-of-care (POC) diagnosis (8, 9). It is encouraging to note that another group has successfully migrated manual real-time PCR assays to multiple sample-to-result platforms, such as the BD Max, ELITe InGenius (ELITechGroup), and Aries (Luminex) simultaneously (10). In summary, the development and validation of the rapid and automated sample-to-result C. auris BD Max open system assay would allow for the wider adaptability and availability of surveillance testing at front-line laboratories.