Comparative evaluation of three dengue duo rapid test kits to detect NS1, IgM, and IgG associated with acute dengue in children in Myanmar

Dengue is an increasing public health concern worldwide and requires efficient laboratory diagnostics. We evaluated three commercially available dengue rapid diagnostic tests—the Humasis Dengue Combo NS1 & IgG/IgM (Humasis, Korea), SD Bioline Dengue Duo NS1 Ag & IgG/IgM (SD Bioline, Korea), and CareUS Dengue Combo NS1 and IgM/IgG kits (WellsBio, Korea)—and compared them to reference immunoglobulin M (IgM) or immunoglobulin G (IgG) ELISAs and quantitative reverse transcription polymerase chain reaction (qRT-PCR) assays. In total, 109 dengue-positive samples from children with acute symptomatic dengue and 63 dengue-negative samples from febrile and asymptomatic individuals were collected. For the nonstructural 1 protein (NS1) Ag test, the sensitivity and specificity were in the following order: CareUS (79.82 and 100%), Humasis (63.30 and 100%), and SD Bioline (48.62 and 100%). For IgM and IgG, CareUS had the highest sensitivities and specificities (89.91 and 100%; 82.57 and 100%, respectively), followed by SD Bioline (60.55 and 100%, 77.98 and 100%, respectively), and Humasis (51.38 and 98.21%, 72.48 and 95.24%, respectively). The IgM kits were more sensitive than the NS1 Ag or IgG kits; however, combining NS1 Ag and IgM reduced the number of missed cases. Therefore, the NS1 Ag plus IgM dengue kits increase the accuracy of the results. In our study, the CareUS Dengue Combo NS1 and IgM/IgG kit showed higher accuracy in performance with reference to qRT-PCR and ELISA results.


Introduction
Dengue fever is caused by four serotypes of dengue virus (DENV)-DENV1, DENV2, DENV3, and DENV4-and is transmitted mainly by Aedes aegypti. Since the end of the 20th

Subjects
Blood samples were collected from patients with suspected dengue fever at the Yankin Children Hospital in Yangon in Myanmar from October 2015 to August 2016. Clinical diagnoses of dengue were categorized as dengue, dengue with warning signs, or severe dengue, according to the WHO guidelines [14]. A total of 189 samples were collected, of which 172 were included in this study (Fig 1). The serum panel consisted of 172 serum samples, including 102 cases confirmed to be positive for DENV infection by qRT-PCR and 7 cases confirmed to be positive for DENV infection based on IgM or IgM/IgG ELISAs. Additionally, the panel included 63 serum samples collected from patients with other illnesses that tested negative for DENV by qRT-PCR, or who tested negative for dengue by IgM and IgG ELISAs.  Evaluation of three dengue duo rapid test This protocol was designed based on in-house real-time RT-PCR and IgM/IgG ELISA. Dengue-positive and -negative cases were used for the comparative evaluation of the performances of three rapid diagnostic (RDT) kits: Humasis Dengue Combo NS1 and IgM/IgG (Humasis, Anyang, Korea), SD Bioline Dengue Duo NS1 and IgM/IgG (Standard Diagnostic, Yongin, Korea), and CareUS Dengue Combo NS1 & IgM/IgG (Wellsbio, Seoul, Korea).

qRT-PCR
The presence of dengue virus and the corresponding serotype was confirmed by qRT-PCR and sequencing. The theoretical specificity of the system was investigated using BLAST against the NCBI nucleotide database. RNA was extracted from the serum sample obtained from each patient with clinically diagnosed with dengue fever using the QIAamp Viral RNA kit (Qiagen, Germany) and used as template for qRT-PCR using a DiaStar OneStep Multiplex qRT-PCR Kit (Solgent, Korea). Newly designed DENV 1-4 primers and probes (S1 Table) were used in this study. The detection limit of the DENV 1-4 primers in qRT-PCR is shown in S2 Table. Amplification and detection were performed using a CFX-96 instrument (BioRad, USA). Thermocycling parameters were as follows: reverse transcription at 50˚C for 30 min, inactivation at 95˚C for 15 min, 40 cycles of fluorescence detection at 95˚C for 40 s, and a final annealing at 62.7˚C for 1 min. Samples were considered dengue-positive if target amplification occurred within 40 cycles. All PCR products were sequenced for DENV serotype confirmation.

Dengue IgM/IgG ELISA
The blood and serum samples were tested for dengue IgM/IgG using a Dengue IgG and IgM ELISA (MyBioSource Inc, San Diego, CA, USA) according to the manufacturer's instructions. Briefly, dengue serum samples were diluted 1:20 (IgG) and 1:1000 (IgM) in dilution buffer, and then 100 μL of the mixture was added to each well. After incubating for 10 min, the wells were washed and 100 μL of enzyme conjugate (anti-dengue IgM or IgG antibody labeled with peroxidase) was added to each well. After incubating for 10 min, the plates were washed and the conjugated complex was visualized by adding 100 μL of TMB solution to the wells. After incubating for 10 min, the reaction was stopped by adding 100 μL of stop solution, and the plate was then read at 450 nm using a microtiter plate reader. Positive determinations were made against the provided IgM and IgG reference sera (cutoff calibrators).

Humasis Dengue Combo NS1 and IgM/IgG
The Dengue Combo NS1 Ag and IgG/IgM Ab Rapid Test is a qualitative membrane-based immunoassay for the detection of NS1 antigen and IgG/IgM antibodies in human serum/ plasma/whole blood. The test kit consists of one device for detecting the dengue NS1 antigen, and a second device for the differential detection of dengue IgG/IgM antibodies. All tests were performed in accordance with the manufacturer's instructions. To detect the dengue NS1 antigen, 100 μL of whole blood, or serum/plasma sample, was added to the sample well of the Humasis Dengue NS1 device. Similarly, for the detection of dengue IgG/IgM, 10 μL of whole blood, or serum/plasma sample, was added to the sample well, followed by the addition of 3 to 4 drops (100-120 μL) of assay buffer to the round assay buffer well. All data were obtained after 15-20 min. Each cassette included a control and a test line, where the appearance of both the control and the test line indicated a positive result, and the appearance of the control line alone indicated a negative result.

SD Bioline Dengue Duo NS1 Ag and IgM/IgG test
The SD Bioline Dengue Duo NS1 Ag and IgG/IgM test kit is a one-step immunochromatographic assay designed to detect both dengue virus NS1 antigen and differential IgM/IgG antibodies to detect the presence of dengue virus in human whole blood, serum, and plasma. The SD Bioline Dengue Duo rapid test contains two tests in one device, where the left and right sides are dengue NS1 antigen and IgG/IgM tests, respectively. All tests in this study were carried out in accordance with the manufacturer's instructions. To detect the dengue NS1 antigen, 100 μL of the test sample was added to the left sample well and the results were read at 15-20 min. To detect dengue IgG/IgM, 10 μL of test sample was added to the right sample well, and 4 drops (90-120 μL) of assay diluent were added to the round assay diluent well, and the results were read 15-20 min later. Each test included a control and a test line, where the appearance of both the control and test line indicated a positive result, while the appearance of only the control line indicated a negative result.

CareUS Dengue Combo NS1 and IgM/IgG
CareUS Dengue Combo NS1 and IgM/IgG tests (WellsBio, Seoul, Korea) are designed to diagnose dengue fever in acute to convalescent patients and consist of a dengue NS1 antigen test strip and an anti-dengue immune IgM/IgG test strip in one device. All tests were carried out in accordance with the manufacturer's instructions. To detect the dengue NS1 antigen, 80 μL of the test sample was added to the left sample well, 1 drop (~40 μL) of the analytical diluent was then added to the same well, and the results were read at 20 min. To detect dengue IgG/IgM, 20 μL of test sample was added to the right sample well, followed by the addition of 1 drop (40 μL) of analytical diluent to the same well, and the results were read at 15-20 min. Each test included control and test lines, where the appearance of both the control and test line indicated a positive result, and the appearance of only the control line indicated a negative result.

Data analysis
Data were tabulated and analyzed using Microsoft Excel (Microsoft Inc., Redmond, WA, USA). Confidence intervals for sensitivity and specificity were set at 95% and a two-tailed p < 0.05 indicated significance. Statistical analysis was performed with Stata v13 (StataCorp, Texas, USA). The sensitivity and specificity for the assays were calculated based on "true positive dengue samples" (PCR and IgM & IgG positives) using a diagnostic test evaluation calculator program (https://www.medcalc.org/calc/diagnostic_test.php) and the Free Statistics Calculators program (https://www.danielsoper.com/statcalc/calculator.aspx?id=29).

Results
A total of 220 samples were assessed in this study. For the dengue-positive cases, there were 109 samples from children with acute symptomatic dengue confirmed by RT-PCR, IgM, and IgM/IgG ELISA (Fig 1). The patient included an almost equal ratio of males (52.29%) and females (47.71%). The mean age of the patients was 7.5 years (range 1-14 years). Acute phase samples from patients with suspected dengue were collected after 3 to 7 days of fever. To distinguish between primary and secondary dengue virus infections, we used a cutoff IgM/IgG ratio of 1.2, which was consistent with the published cutoff ratio range [15,16]. There were 22 primary (20.2%) and 87 secondary infections (79.8%) among the dengue-positive cases. The serotype distribution of the 102 RT-PCR-based positive samples were DENV1 (77; 75.5%), DENV2 (23; 22.5%), and DENV4 (2; 2%). According to WHO guidelines, dengue-positive samples were classified as dengue (19; 17.4%), warning signs (63; 57.7%), or severe dengue (27; 24.7%). The dengue-negative cases yielded 63 samples from children with febrile non-dengue infections. All these samples were negative for dengue based on qRT-PCR and an anti-dengue IgM and IgG ELISA (Fig 1).
The left, middle, and right panels show NS1, IgM, and IgG RDT tests, respectively. The numbers in parenthesis below the x-axes indicate the number of patients with dengue according to the dengue classifications based on WHO guidelines.

Clinical performances of three RDT kits for acute dengue infection
The sensitivity and specificity for the NS1 antigen, IgM, and IgG for the three RDTs tested (Humasis, SD Bioline and CareUS) are shown in Table 1

Clinical performance of the three dengue RDTs in patients with primary and secondary serological profiles
Since the concentrations of NS1 and DENV IgM/IgG antibodies in the blood depends on whether the infection is primary or secondary, the sensitivities of three DENV RDT kits were determined ). When we analyzed the sensitivity of all biomarkers according to the classification as a primary (NS1/IgM/IgG -) or a secondary infection (NS1/IgM/IgG + ), the Humasis RDT kit had similar sensitivities for primary (63.64%) and secondary infections (65.52%). However, the sensitivities of the SD Bioline (68.97%) and CareUS (95.40%) RDT kits for secondary infections were higher than for primary infections (SD Bioline, 50.00%; CareUS, 59.09%). Among the three RDT kits, the CareUS RDT kit was the only kit displaying a statistically significant difference between primary and secondary DENV infections (p < 0.05).

Clinical performance of the three RDT kits according to the day after the onset of fever
When comparing the sensitivities of the three NS1 RDT kits for primary and secondary infections according to day after onset of fever, all three NS1 RDT kits had continuously high sensitivities for primary infection during the entire febrile period, whereas the sensitivities of all three NS1 RDT kits for secondary infections gradually decreased with increasing days after the onset of fever (Fig 2A). The CareUS NS1 RDT kit had the highest sensitivity for both primary and secondary infections during the entire febrile period. The sensitivities of all three IgM RDT kits for primary infections drastically increased at the 5th day of fever, whereas the sensitivities of all three kits for secondary infections remained continuously high and did not change significantly during the entire febrile period (Fig 2B). For primary infections, the three IgG RDT kits had very low sensitivities during the congested pattern during the entire febrile period. In contrast, the sensitivities of all three IgG RDT kits for secondary infections remained continuously high during the entire febrile period (Fig 2C).

Real-time PCR threshold cycle (Ct) levels of dengue RDT kits for dengue-positive cases
For all the NS1 antigen tests, the Ct values for NS1 antigen-positive cases were statistically significantly lower (p < 0.05, except CareUS with a p = 0.0611) than for NS1 antigen-negative cases (Table 3). Among the three NS1 kits, the Ct values (32.18 ± 8.70) for the NS1 antigenpositive cases (N = 87) tested by the CareUS kit were higher than those of the other RDT kits.
In the case of the IgM tests, there was no statistically significant difference (p < 0.05) between IgM-positive and IgM-negative cases for all the test kits, despite the Ct values of the IgM-positive cases being lower than those of the IgM-negative cases. Meanwhile, the Ct values of the IgG positive cases for all the IgG tests were statistically significantly (p < 0.05, except for SD Bioline with a p = 0.0632) higher than those of the IgG-negative cases.

Comparison of the sensitivities of the three RDT kits according to dengue serotype
The sensitivities of the three RDT kits for the dengue serotypes were compared and evaluated (Fig 4). For the NS1 test with respect to DENV1 and DENV2 cases, CareUS had the highest

Discussion
Dengue fever is caused by the dengue virus and is a tropical disease transmitted by mosquitoes [17,18]. Secondary infections can lead to life-threatening dengue hemorrhagic fever or dengue shock syndrome. As stated by the WHO for the Southeast Asia and Western Pacific region, about 1.8 billion people are at risk of dengue, putting a significant economic and disease burden on these countries [19,20]. Myanmar has seen an increase in dengue fever over the past decade, and is now classified as a high-risk dengue country in the Asia-Pacific region [21]. At  [8] were derived from a small sample of 36 patients. In fact, the distribution of dengue type over the entire 2013-2015 period was similar to our results. In this study, we compared and evaluated the diagnostic usefulness of three RDTs, Humasis, SD Bioline, and CareUS dengue kits based on dengue diagnosis criteria using RT-PCR and single serum for IgM/IgG. When the SD Bioline NS1 and IgM RDT kit results were compared with those from an earlier study [23], the sensitivity (80.73%) and specificity (100.00%) of the SD Duo NS1/IgM in this study were similar to those (88.65 and 98.75%) in recent studies [23]. The Humasis and CareUS DENV kits were the first tested in this study. Of the three DENV RDT kits in this study, CareUS had the highest sensitivity for NS1 (79.82%), IgM (89.91%), NS1/IgM (96.33%), and IgG (82.57%). Interestingly, the Humasis kit exhibited a higher sensitivity for NS1 than IgM, whereas the SD Bioline kit had a higher sensitivity for IgM than NS1. Combining the NS1 and IgM test results yielded the highest sensitivities for all RDT kits. The sensitivities of the SD Bioline and Humasis kits were markedly increased when the NS1 and IgM test results were combined.
It is important to distinguish between primary and secondary dengue fever when treating patients, because secondary infections can lead to dengue hemorrhagic fever and dengue shock syndrome with serious complications [24]. Primary and secondary infections have different NS1 antigen and IgM/IgG antibody profiles [25]. It has been shown in longitudinal studies on patients with dengue that patients remain NS1 antigen-positive after the DENV RNA amplicons disappear [26][27][28], probably because of the longer half-life of the NS1 protein.
However, it has been reported that the NS1 detection test had low sensitivity for secondary dengue infections due to the immune complex formation of the NS1 antigen with pre-existing antibodies [29]. Indeed, the sensitivities of the three NS1 RDT kits in this study also gradually decreased for secondary dengue infections as the length of illness increased for all test kits, whereas the detection of NS1 in primary dengue remained high during the entire febrile period. However, the three NS1 RDT kits had high sensitivities for both primary and secondary DENV infections within 4 days of onset of fever. For the IgM tests, the sensitivities of all three IgM RDT kits for primary infections drastically increased after 5 days of illness, whereas the sensitivity for secondary infections remained high without significantly changing during the entire febrile period, suggesting that NS1 and DENV IgM detect DENV during different periods of the DENV primary infection. In contrast, the three IgG RDT kits had very low sensitivities for the congested pattern in primary DENV infections during the entire febrile period, whereas the sensitivities of all three IgG RDT kits for secondary infections remained high during the entire febrile period. In particular, when analyzing the sensitivities for all biomarkers according to primary (NS1/IgM/IgG -) and secondary infections (NS1/IgM/IgG + ), the CareUS kit had statistically different (p <0.05) sensitivities of 59.09 and 95.40%, respectively. SD IgG kit showed no cross-reactivity with Chikungunya virus infected serum. All three NS1/ IgM kits showed no cross reactivity with Chikungunya virus infected serum. However, in cross-reactivity test of three RDT kits against Chikungunya virus infected serum samples (n = 15) among other flavivirus, Humasis and CareUS IgG test showed cross-reactivities 80% and 73.3%, respectively (S2 Table). All three NS1/IgM kits showed no cross-reactivity with Chikungunya virus infected serum samples. Although we did not perform cross-reactivity testing of the three test kits for other flavivirus samples such as Japanese encephalitis virus (JEV), or Zika virus (ZIKV), which are common in southeast Asia, Humasis and CareUS IgG test might have cross-reactive activity with other flavivirus infected samples. Therefore, for confirmation of dengue infection only IgG is not recommended without additional NS1 Ag and IgM antibody status, and it should be checked whether another flavivirus is infected.
The threshold cycle (Ct) values reflect the overall amount of viral load in each clinical specimen, and the relative viral load can be estimated from the Ct value of each specimen. For specimens with low Ct values, the three RDT kits showed comparable sensitivities. However, for specimens with higher Ct values, the NS1 antigen and IgM of CareUS kit only detected Dengue viruses. These findings suggest that the CareUS kit detects lower viral loads more precisely than do the two other RDT kits.
The limitations of this study are that there were not enough DENV4 samples to analyze test performance and DENV3 samples were not collected. The results for adult samples may also be different, since only samples from children were used in this study. The sensitivity and specificity results in this study may be difference with the results classified by the standard dengue diagnosis (RT-PCR and paired serum for IgM/IgG) recommended by WHO, because dengue and non-dengue cases in this study were classified by using RT-PCR and single serum for IgM/IgG. In addition, the sensitivity of the three tested RDT kits may have decreased slightly because patient blood was stored at -80˚C and shipped to Korea from Myanmar for testing.
In conclusion, our study suggests that testing for both dengue NS1 and IgM/IgG is more effective for diagnosis than either alone. RDTs are simple, easy to perform, inexpensive, sensitive, and require no sophisticated protocols, and, therefore can be used in any location. Among the kits evaluated, the CareUS RDT kit was the most efficient when confirming dengue infections by capturing the NS1 antigen and detecting DENV-IgG/IgM from infected patients. Moreover, it was the most convenient to use (as the results could be obtained in 15 min), easy to perform, and it did not require special laboratory equipment.
Supporting information S1 Table. In-house primer and probe sequences used for dengue detection. (DOCX) S2