Comparison of Two Commercial Real-Time PCR Assays for Detection of Dengue Virus in Patient Serum Samples

Abstract

This study evaluated and compared the performance of two real-time PCR assays using nested reverse transcription (RT)-PCR as the reference method. Among 117 nested RT-PCR-positive cases, the abTES DEN 5 qPCR kit detected 97.4% of dengue virus (DENV) infections, while the innuDETECT Dengue TwoStep assay did so for 44.4%. Sensitivity varied by infecting serotype and the stage of infection. The abTES kit has the potential to replace nested RT-PCR for the rapid diagnosis of DENV infection.

TEXT

Dengue virus (DENV) infection is a major public health problem in Thailand. From January to December 2013, the number of reported cases of DENV infection was >150,934, and there were 133 deaths (1). DENV belongs to the Flavivirus genus in the Flaviviridae family. Its genome consists of three structures, including a core protein, a precursor membrane protein (prM), and an envelope protein, as well as seven nonstructural proteins (NS1, NS2a, NS2b, NS3, NS4a, NS4b, and NS5) (2). It is often difficult to distinguish DENV infection from influenza virus or other viral infections using signs and symptoms alone. Screening tests, such as the dengue NS1 antigen (Ag) enzyme immunoassay (EIA) and the IgM/IgG EIA, are utilized; however, the sensitivities of these tests are limited, partly due to cross-reactions among species of Flavivirus, and the IgM/IgG EIA is suitable for diagnosis at later stages. Moreover, viral cultures are mainly performed in special laboratories due to their limitations, such as being labor intensive and time-consuming (3–5). Presently, molecular techniques for confirming DENV infection can be used to differentiate among DENV serotypes 1, 2, 3, and 4 (DENV-1 to -4). In Thailand, nested reverse transcription-PCR (RT-PCR) is used by the Ministry of Public Health as the reference method for confirming DENV infection. Although this method provides high sensitivity and specificity, it is a two-stage assay that is quite time-consuming (6). Real-time PCR assays have been implemented in many laboratories to detect DENV infection (7, 8). This study aimed to evaluate the sensitivities and specificities of two commercial real-time PCR assays, the abTES DEN 5 qPCR (abTES) kit and the innuDETECT Dengue TwoStep assay (innuDETECT), for the detection of DENV infection.

In total, 125 serum samples were obtained from patients suspected of having DENV infection at Phramongkutklao Hospital (Bangkok, Thailand) from January 2001 to December 2012. This study was approved by the Institutional Review Board of the Royal Thai Army Medical Department (Bangkok) and the Committee on Human Rights Related to Research Involving Human Subjects of Thammasat University (Pathumtani, Thailand). Results from reference dengue assays (nested RT-PCR and IgM/IgG enzyme-linked immunosorbent assay [ELISA]) performed at the Armed Forces Research Institute of Medical Sciences (AFRIMS) were available as part of ongoing public health service testing for Phramongkutklao Hospital, and nested RT-PCR was used as the reference method for confirming DENV infection.

Total viral RNA was extracted from 150 μl of each serum sample using an innuPREP viral RNA kit (Analytik Jena, Jena, Germany) according to the manufacturer's protocol (9). All the samples were tested for DENV infection according to the abTES (AITbiotech, Singapore) (10) and innuDETECT (Analytik Jena) protocols (11). Real-time PCR was performed using the CFX96 real-time thermocycler (Bio-Rad Laboratories, Hercules, CA, USA).

The data were analyzed using STATA/MP version 12 (Stata Corp LP). Exploratory data analysis was applied to evaluate the sensitivities, specificities, positive predictive values (PPV), and negative predictive values (NPV) of the two real-time PCR assays using nested RT-PCR as the reference method. McNemar's test was used to determine agreement across the two real-time PCR assays in test positivity. P values of <0.05 were considered statistically significant. The degree of concordance of the results between the reference method and other methods was calculated using the κ index and 95% confidence intervals (95% CI). The κ value was interpreted as follows: a κ value of 0.00 to 0.20 indicates poor agreement, a κ value of 0.21 to 0.40, fair agreement, a κ value of 0.41 to 0.60, moderate agreement, a κ value of 0.61 to 0.80, good agreement, and a κ value of 0.81 to 1.00, near-perfect agreement (12).

Altogether, 125 serum samples from patients suspected of having DENV infection were included; 117 samples were positive by DENV nested RT-PCR and IgM/IgG ELISA, and 8 samples were negative by both tests. The 117 DENV-positive samples were additionally classified as indicating primary or secondary DENV infection using the IgM/IgG ELISA results. Nested RT-PCR was used to differentiate DENV into four serotypes, DENV-1 (n = 32), DENV-2 (n = 27), DENV-3 (n = 34), and DENV-4 (n = 24). Of these, 59 samples were positive by RT-PCR and the first round of nested PCR (RT-PCR positive). Additionally, 58 samples were positive only after the second round of nested PCR (RT-PCR negative).

The abTES assay showed a sensitivity of 97.4%, correctly identifying 114 out of 117 samples as positive. For these 114 samples, 59/59 (100%) were positive in the RT-PCR-positive samples, and 55/58 (95%) were positive in the RT-PCR-negative samples. The innuDETECT assay showed a sensitivity of 44.4%, correctly identifying 52 out of 117 samples as positive. Regarding these 52 samples, 45/59 (76.3%) were positive in the RT-PCR-positive samples, and 7/58 (12.1%) were positive in the RT-PCR-negative samples. Eight control serum samples were negative by both assays, resulting in a specificity of 100%. The κ index for DENV infection by the abTES assay showed near-perfect agreement (κ = 0.83), while the innuDETECT assay showed poor agreement (κ = 0.09) compared with the reference method (Table 1). Both real-time PCR assays had statistically significant differences (P < 0.001) in sensitivity. The sensitivities and specificities of both assays by serotype (DENV-1 to -4) and the stage of infection are shown in Tables 1 and 2, respectively.

TABLE 1.

Analytical analysis: Summary of sensitivities, specificities, PPV, NPV, and κ values between nested RT-PCR (reference method) and two commercial kits according to DENV serotype

Test and serotypea	Sensitivity (95% CI) (%)	Specificity (95% CI) (%)	PPV (95% CI) (%)	NPV (95% CI) (%)	κ
innuDETECT Dengue TwoStep assay
DENV-1	53.1 (34.7–70.9)	100.0 (63.1–100)	100.0 (80.5–100)	34.8 (16.4–57.3)	0.31
DENV-2	40.7 (22.4–61.2)	100.0 (63.1–100)	100.0 (71.5–100)	33.3 (15.6–55.3)	0.23
DENV-3	38.2 (22.2–56.4)	100.0 (63.1–100)	100.0 (75.3–100)	27.6 (12.7–47.2)	0.19
DENV-4	45.8 (25.6–67.2)	100.0 (63.1–100)	100.0 (71.5–100)	38.1 (18.1–61.6)	0.29
All serotypes	44.4 (35.3–53.9)	100.0 (63.1–100)	100.0 (93.2–100)	11.0 (4.85–20.5)	0.09
abTES DEN 5 qPCR
DENV-1	100.0 (89.1–100)	100.0 (63.1–100)	100.0 (89.1–100)	100.0 (63.1–100)	1.00
DENV-2	100.0 (87.2–100)	100.0 (63.1–100)	100.0 (87.2–100)	100.0 (63.1–100)	1.00
DENV-3	94.1 (80.3–99.3)	100.0 (63.1–100)	100.0 (89.1–100)	80.0 (44.4–97.5)	0.85
DENV-4	95.8 (78.9–99.9)	100.0 (63.1–100)	100.0 (85.2–100)	88.9 (51.8–99.7)	0.92
All serotypes	97.4 (92.7–99.5)	100.0 (63.1–100)	100.0 (96.8–100)	72.7 (39.0–94.0)	0.83

^aSerotypes were confirmed by nested RT-PCR (reference method).

TABLE 2.

Analytical analysis: Summary of sensitivities, specificities, PPV, NPV and κ values between nested RT-PCR (reference method) and commercial kits according to stage of infection

Test and stage of DENV infectiona	Sensitivity (95% CI) (%)	Specificity (95% CI) (%)	PPV (95% CI) (%)	NPV (95% CI) (%)	κ
innuDETECT Dengue TwoStep assay
Primary	50.0 (30.6–69.4)	100.0 (63.1–100)	100.0 (76.8–100)	36.4 (17.2–59.3)	0.31
Secondary	42.7 (32.3–53.6)	100.0 (63.1–100)	100.0 (90.7–100)	13.6 (6.04–25)	0.11
Primary plus secondary	44.4 (35.3–53.9)	100.0 (63.1–100)	100.0 (93.2–100)	11.0 (4.85–20.5)	0.09
abTES DEN 5 qPCR
Primary	100.0 (87.7–100)	100.0 (63.1–100)	100.0 (87.7–100)	100.0 (63.1–100)	1.00
Secondary	96.6 (90.5–99.3)	100.0 (63.1–100)	100.0 (95.8–100)	72.7 (39.0–94.0)	0.83
Primary plus secondary	97.4 (92.7–99.5)	100.0 (63.1–100)	100.0 (96.8–100)	72.7 (39.0–94.0)	0.83

^aStages of DENV infection were classified by IgM/IgG ELISA.

Nested RT-PCR is routinely used as a confirmatory test for DENV infection; however, the test is limited by being time-consuming. Currently, real-time PCR assays are implemented to detect DENV infection (7, 13, 14). In this study, the sensitivities and specificities of two real-time PCR assays were evaluated using nested RT-PCR as the reference method. False-negative results were found in both real-time PCR assays, caused by a low viral load in the first round of nested PCR that was RT-PCR negative and by insufficiency of the cDNA template in the second round of nested PCR.

Timely diagnosis of DENV infection is recommended to reduce life-threatening risks, especially in cases of negative NS1 Ag and IgM/IgG antibody results. Although the total processing time for nested RT-PCR is 6.5 h, the results are usually not available to clinicians until 2 days after blood collection. Real-time PCR assays can be less time-consuming than nested RT-PCR assays and are simple and suitable for the rapid confirmation of DENV infection in reference laboratories.

In conclusion, the abTES assay is more sensitive than the innuDETECT assay and is less time-consuming than nested RT-PCR, providing rapid diagnosis of DENV infection that can be helpful in urgent clinical situations.