Validation and Application of a Commercial Quantitative Real-Time Reverse Transcriptase-PCR Assay in Investigation of a Large Dengue Virus Outbreak in Southern Taiwan

Background Accurate, rapid, and early diagnosis of dengue virus (DENV) infections is essential for optimal clinical care. Here, we evaluated the efficacy of the quantitative real-time PCR (qRT-PCR)-LightMix dengue virus EC kit for DENV detection using samples from a dengue outbreak in Taiwan in 2015. Methods Sera from patients with suspected DENV infection were analyzed and compared using the LightMix kit, a Dengue NS1 Ag + Ab Combo kit for detection of NS1 antigen and DENV-specific IgM and IgG antibodies, and an “in-house” qualitative DENV-specific RT-PCR assay. Results A total of 8,989, 8,954, and 1581 samples were subjected to NS1 antigen detection, IgM and IgG detection, and LightMix assays, respectively. The LightMix assay yielded a linear curve for viral loads (VL) between 102 and 106 copies/reaction, and the minimum detection limits for DENV serotype 1 (DENV1) and DENV2, DENV3, and DENV4 were 1, 10, and 100 focus forming units (FFU)/mL, respectively. There was 88.9% concordance between the results obtained using the NS1 antigen combo kit and by LightMix analysis, and the diagnostic sensitivity and specificity of the two methods were 89.4 and 100%, and 84.7 and 100%, respectively. Notably, fatal cases were attributed to DENV2 infection, and 79.5% (27/34) of these cases occurred in patients ≥ 71 years of age. Among these older patients, 82.3% (14/17) were NS1/IgM/IgG (+/-/-), exhibiting VLs between 106–109 copies/mL, which was markedly higher than the rate observed in the other age groups. Conclusions The LightMix assay was effective for early diagnosis of DENV infection. Our data indicate that high VLs during primary infection in elderly patients may be a positive predictor for severe illness, and may contribute to high mortality rates.


Introduction
Between August and November 2015, there were 112 deaths and 22,563 confirmed cases (mortality rate: 0.5%) of DENV2 in a large outbreak in Tainan, Taiwan, as reported by the Taiwan Centers for Disease Control (CDC) [28]. Here, we validate the efficacy of the LightMix dengue virus EC kit, a newly implemented commercial quantitative real-time (qRT)-PCR assay, for the detection of DENV infection in the outbreak.

Ethics statement
This retrospective study was conducted in Tainan, Taiwan. Approval for the study was obtained from the Institutional Review Board (IRB) of National Cheng Kung University Hospital (No. B-ER-104-228). The demographic and clinical information for the patients were delinked prior to analysis. Thus, informed consent was not obtained from patients prior to the study.

Clinical samples
Serum samples from patients with suspected DENV infection were collected at Clinical Virology Laboratory of National Cheng Kung University Hospital (NCKUH) between July and November 2015. Almost all samples were screened via a rapid combo test for antigen and antibody detection for clinical diagnosis; the LightMix dengue virus EC qRT-PCR and serotyping assays were also performed on portions of these samples, per clinical requirements. Patients were categorized as having mild, severe, or fatal cases of dengue according to 2009 World Health Organization (WHO) criteria for severity. For this study, severe dengue cases were considered those that fulfilled one of the following 2009 WHO criteria: severe plasma leakage, severe bleeding, or severe organ involvement.

NS1 antigen, and IgM and IgG antibody detection
Sera from patients with suspected DENV infection were screened using the one-step immunochromatographic Dengue DuoDengue NS1 Ag + Ab Combo assay (SD BIOLINE, Yongin, Korea). This rapid assay contains two test devices, which can simultaneously detect the levels of NS1 antigen, and IgM and IgG in a sample, respectively, within 15 minutes.
RNA extraction, reverse transcription, qRT-PCR assay, and serotyping Viral RNA was extracted from serum samples and an extraction control sample using a QAIamp viral RNA mini kit (Qiagen, Venlo, Netherlands) or the automated extraction system LabTurbo Virus mini Kit in LabTurbo 48 Compact System (Taigen Bioscience Corp., Taipei, Taiwan). qRT-PCR analysis of viral loads was performed using lightMix dengue virus EC kit (qRT-PCR; TIB Molbiol, Berlin, Germany), which is capable of identifying all four dengue serotypes. cDNA was generated from the RNA samples using a FirstStrand cDNA Synthesis kit (Roche, Basel, Switzerland) and a DENV-specific primer included in the LightMix dengue virus EC kit. The qRT-PCR assay was then performed in a LightCycler 2.0 or LightCycler 480 II device (Roche), according to the manufacturer's instructions; briefly, cDNA was added to a PCR reagent mix containing PCR grade water, the LightMix dengue virus EC kit reagents (including the primers and probes of DENV and the external control), and the mastermix of Roche LightCycler R FastStart DNA Master HybProbe. PCR conditions were as described in the instructions for the LightMix dengue virus EC kit. The amplicon corresponding to the 3'-noncoding region (NCR; 164-187 bp) of DENV was detected using LightCycler Red 640-labeled hybridization probes, while the extraction control (349 bp) derived from Lambda DNA was detected using fluorescence resonance energy transfer (FRET) probes labeled with LightCycler Red 690 (channel 705). The LightMix dengue virus EC kit provides cloned dengue DNA at concentrations of 10 1 to 10 6 copies/reaction as standards. The cycle number of the Crossing Point (Cp) of each sample was calculated automatically by the Second Derivative Maximum method (Automated (F" max)), with LightCycler's software. A logarithmic transformation was applied to all resulting data. The amount of virus per sample (viral load) was reported in copies/reaction, as automatically generated by the LightCycler software.
Serotyping was conducted by an in-house qualitative RT-PCR method, as reported by Tanaka et al. [29].
The accuracy, precision, and analytical sensitivity of the LightMix dengue virus EC kit were evaluated using the reference strains listed above or triplicates of serially diluted plasmid DNA standards (10 1 to 10 6 copies/reaction), which were provided with LightMix dengue virus EC kit. Within-run samples were tested in triplicate during the same run, while positive values from another two runs were used for evaluating between-run precision, which represent the coefficient of variation (CV). In parallel, the analytical specificity of the assay was tested using clinical samples containing various viral species. Lastly, the diagnostic sensitivity and specificity of the assays were determined by comparing the results obtained from clinical specimens collected from suspected DENV-infected patients with those obtained by the qRT-PCR assay, inhouse qualitative dengue RT-PCR, and NS1 antigen assays.

Statistical analysis
Frequency distributions for analyses of viral load (VL) levels and the results of NS1 antigen, and IgM and IgG assays were compared by χ 2 test. P values < 0.05 were considered statistically significant. All analyses were conducted using SPSS 17.0 software (SPSS Statistics, Inc., Chicago, IL, USA).

Description of the dengue virus outbreak
A total of 8,989, 8,954, and 1,581 samples were subjected to NS1 antigen detection, IgM and IgG detection, and qRT-PCR assay analysis, respectively. Of 8,954 samples were screened via a rapid combo test for antigen and antibody detection for clinical diagnosis. Meanwhile, 1,581 specimens collected from patients with suspected DENV infections were simultaneously subjected to NS1 antigen, IgM and IgG, and qRT-PCR assay analyses per clinical requirements. Of these samples, 41.8% (3755/8,989), 11.2% (999/8,954), 6.9% (618/8,954), and 40.2% (619/ 1,538) tested positive for NS1, IgM, and IgG, and tested positive for DENV in the qRT-PCR assay, respectively. As shown in Fig 1, the peak of the outbreak occurred between the 5 th week of August and the 4 th week of September (Fig 1).
Of the 44 samples collected from patients with fatal cases (25 female, 19 male) of dengue, 34 tested positive for DENV with the LightMix dengue EC kit, which targets the conserved region of DENV serotypes 1-4. Notably, each of these 34 samples, as well as 67 of the 85 samples obtained from patients diagnosed with severe cases (33 female, 52 male), tested positive for DENV2 by our in-house dengue RT-PCR serotyping assay. Conversely, none of these samples was positive for DENV1, DENV3, or DENV4. The age of the patients associated with fatal disease ranged from 41 years to 91 years, but 42 (95.4%) of the patients were older than 60 years of age.

Validation of the LightMix Dengue virus EC qRT-PCR assay
To validate the LightMix Dengue virus EC kit, we evaluated the precision, analytical sensitivity, analytical specificity, clinical sensitivity, and clinical specificity, of assay for diagnosing DENV infection. The between-run and within-run precisions, and the linearity of the assay were evaluated using dilutions of the plasmid DNA standard (10 1 to 10 6 copies/reaction) provided with the LightMix kit. While the CVs for the within-run precision when using 10 2 to 10 6 plasmid copies/reaction were 0.29%-7.24%, the CV when using 10 1 copies/reaction was markedly higher (40.29%; Table 1). Meanwhile, the between-run precision values were 1.25%, 1.16%, and 1.22% for 10 2 , 10 4 , and 10 5 plasmid copies/reaction, respectively (Table 1). Linearity, which indicates the reportable quantifiable range of the assay, was observed between 10 2 and 10 6 copies/reaction (R 2 = 0.99; Fig 2), and the minimum quantification limit (QL) was 10 2 copies/reaction.
For analytical sensitivity testing, quantitative viral reference strains representing the four DENV serotypes were used to spike normal human plasma, adjusted to 10 5 FFU/mL, and 10-fold serially diluted. The limit of detection (LOD) of the LightMix dengue virus EC qRT-PCR assay for the four subtypes was 1 FFU/mL for DENV1 and DENV2, 10 FFU/mL for DENV3, and 100 FFU/mL for DENV4 (Table 2). Each assay was conducted with 10 replicates, and the positive detection rate was 97.5% (39/40 samples; S1 File). that the presence of these 11 other viruses had no effect on the performance of the LightMix dengue virus EC qRT-PCR assay. A total of 199 clinical specimens with suspected dengue virus infection were subsequently evaluated simultaneously by both the LightMix dengue virus EC qRT-PCR assay and another in-house dengue-specific RT-PCR assay. The results obtained from each sample were then compared to a composite reference standard, defined as the detection of DENV in one or both assays. The diagnostic sensitivity of the LightMix dengue virus EC qRT-PCR and in-house RT-PCR assay were 99.4% (179/180) and 72.6% (130/180), respectively.

Discussion
Currently, detection of DENV-specific nucleic acids during the acute phase of infection is widely applied for dengue diagnosis. Indeed, many studies have evaluated the efficacy of realtime RT-PCR assays for the detection of DENV. The results of these previously published reports are summarized in Table 5. The majority of the studies focused on in-house and Table 3  TaqMan assays, as opposed to commercial SYBR and FRET assays. Additionally, the most popular target region used for detection was the 3'-NCR. Meanwhile, the 3'-NCR and NS5 regions, as well as combinations of different genes, have commonly been utilized as target regions for serotyping [14-16, 19-23, 25-27, 30-32]. Although Levi et al. (2007) observed high levels of agreement between the commercial RealArt real-time RT-PCR kit and an in-house multiplex RT-PCR for detection of DENV3 (and occasionally DENV2), the sensitivity of these published methods are highly diverse (3.8-100%) [33]. Moreover, among the 37 laboratories that participated in an international external quality control assessment study to evaluate various denguespecific RT-PCR assays, 80.4% reported a lack of sensitivity for dengue molecular diagnosis [34]. In this study, we detected DENV in 100% (102/102) of the DENV2-positive samples by the LightMix dengue virus EC qRT-PCR assay, compared to only 66.7% (68/102) of the same samples by another in-house qualitative serotyping RT-PCR assay. To date, this is the first report to evaluate a commercially available kit for detection and quantification of DENV via FRET and real-time PCR analyses. A linear curve for the LightMix dengue virus EC qRT-PCR assay was obtained between 10 2 and 10 6 copies/reaction (R 2 = 0.99). Notably, the assay was more sensitive for detection of DENV1 and DENV2 than for DENV3 and DENV4. Such differences in sensitivity for distinct viral serotypes have been reported in the other studies [22,24]. Indeed, these differences were expected because of the free noninfectious virus particles present in the virus stocks, which were quantitated by plaque assay analysis, and because each assay was limited by its primer and probe design. Therefore, we evaluated the diagnostic sensitivity of the assay to validate its use for detection of DENV in specimens derived from infected patients. For these analyses, we compared the results obtained by NS1 antigen, IgM, IgG, and LightMix dengue virus EC qRT-PCR assay analyses of samples collected from suspected DENV-infected patients during an outbreak in Taiwan in 2015. The NS1/IgM/IgG (+/-/-) category was associated with the highest proportion (274/417; 65.7%) of samples exhibiting high VLs (10 6 -10 9 copies/mL). Meanwhile, the highest percentage (88.5%) of specimens with medium VLs (<10 6 copies/mL) was observed in the NS1-positive and IgM and/or IgG-positive category. Notably, the highest ratios (80.7%; 84/  104; j2/(j1 + j2)) of antibody-positive and -negative specimens (83.3%, 60/72; c1/(c1 + c2)) were observed in the NS1/VL (+/-) and NS1/VL (-/+) categories, respectively (Table 3). These data indicate that NS1 antigen can remain within the human sera even after antibody-mediated reduction or clearance of the virus. Consistent with this conclusion, Hernandez et al. (2013) [40] previously demonstrated that while dengue IgM antibody-negative patients exhibit significantly higher VLs than IgM-positive patients, the presence of IgM does not correlate with NS1  [13] found no significant differences in VL between patients with primary and secondary infections. In this study, 26 of the 34 specimens in the NS1/IgM/IgG/VL (+/-/+/+) category (76.4%) had VLs of 10 3 -10 6 copies/mL, indicating the presence of a suspected secondary infection group with medium VLs among the patients examined during this outbreak. In contrast, Laue et al. (1999) reported >5 × 10 6 copies of DENV RNA in samples collected from two suspected dengue secondary infection patients that were IgG-positive but IgM-negative on the first and second day after the onset of illness [30].

Variable Categories for NS1 antigen/IgM/IgG detection [no. of specimens (%)] +/-/-+/+/-+/-/+ +/+/+ -/-/--/+/--/-/+ -/+/+ No. of total specimens
The results of our study suggest that secondary infections associated with medium VL may be modulated by IgG antibody. The correlations between severe dengue disease and high VL, secondary infection, and DENV2 infection have been discussed in multiple studies [8,9,41]. According to data collected by the Taiwan CDC (Centers for Disease Control), 95% of the fatal cases in this outbreak presented with comorbidities such as hypertension (72.9%), diabetes (47.7%), chronic kidney disease (31.8%), heart diseases (17.8%), and cancer (11.2%) [28]. Therefore, we found that high VLs of DENV2 contributed to the high mortality rates observed in elderly patients in primary DENV infections. Patients with comorbidities may also play a role which needs to be examined further. Ahmed et al. (2014) observed maximal sensitivity of the NS1 antigen and real time RT-PCR assays for DENV detection on the second and third day of illness, respectively [11]. The group therefore recommended the application of these two rapid and efficient assays as early diagnostic tests during the acute phase of infection. In this study, there was concordance between the results obtained using these two methods. Furthermore, the diagnostic sensitivity of the two assays was above 80%. Additionally, while a high proportion of the 617 (c0+c+j1+j2) NS1-positive specimens was also VL-positive (513/617, (c0+c)/(c0+c+j1+j2), 83.1%), only 72 of the 964 (72/964, (c1+c2)/(c1+c2+j3), 7.4%) NS1-negative specimens were VL-positive. These data indicate that NS1 antigen detection could be used for rapid diagnostic screening during large outbreaks, while the LightMix dengue virus EC qRT-PCR assay could be employed as a supplemental test for analysis of specimens that test NS1-negative during the acute phase of infection.
There were limitations to our study. Notably, the NS1/IgM/IgG (+/-/-) category contained 20 VL-negative specimens (Table 3), which might indicate the presence of false-negative results owing to variant virus or virus-free in sera. As such, the use of primers or probes to detect additional DENV target regions, and the evaluation of peripheral blood mononuclear cells instead of sera could be considered alternative methods for confirming DENV diagnoses. Accurate, rapid, and early diagnosis of DENV infections could reduce the risk of development of lifethreatening dengue illness, particularly in NS1/IgM/IgG-negative cases. Our results indicate that this LightMix dengue virus EC qRT-PCR assay could be helpful in predicting mortality in primary infected elderly patients, for early diagnosis of DENV infections, and for controlling future dengue outbreaks.