Introduction

Dengue fever (DF) is a major public health problem with 50 million annual cases worldwide. It has been reported in more than 100 countries and continues to spread to previously unaffected regions[1]. DF is caused by dengue viruses (DENV), which consist of four serotypes: DENV-1, DENV-2, DENV-3 and DENV-4. Infection with any serotype usually results in asymptomatic infection or mild non-specific febrile illness, but in a subset of patients, severe disease develops, characterized by a transient capillary leakage syndrome (dengue hemorrhagic fever/ DHF)[2]. This unpredictable disease course and the need to differentiate DF from other causes of fever make an early and sensitive diagnosis of acute dengue virus infection important.

Non-structural 1 protein (NS1) is encoded by the virus and secreted in a soluble form[3]. Several commercial assays to detect dengue NS1 have been developed recently[4]. Advantages of NS1 tests are: NS1 is detected early in disease, several days prior to the appearance of anti-dengue IgM antibodies[5], and the test is inexpensive, easy, and fast. Moreover, NS1 levels early in dengue disease have been correlated with disease severity [6], suggesting that NS1 tests may also have prognostic value.

Previous studies evaluating the diagnostic value of NS1 antigen assays have had varying results [4,7] (Table S1). Several factors may account for this: proportion of primary vs. secondary infections [4,8-11], timing of sample collection[4,7,8,12,13], infecting serotype[4,7,8,13,14], viremia levels [6,13,15,16] and severity of illness[6,7,15]. A low sensitivity has been reported in secondary infections possibly due to immune-complex formation of NS1 and pre-existing antibodies [17]. This may especially be important as the proportion of secondary infections is expected to rise worldwide [18]. Although studies evaluating the NS1 assays were conducted in hyper-endemic areas of Southeast Asia or South America, in several studies, all four dengue serotypes were not represented equally [4,7,8,10,13,15,16,19,20] or evaluated [11,21]. In others, the percentage of primary infections was too high to represent the situation in hyper-endemic areas [9,11,12,15,20]. Additionally, only a few studies evaluated the association between NS1 detection and clinical severity [4,6,7,15,21]. The objectives of the study were: 1) Determine diagnostic performance of the Bio-Rad Platelia Dengue NS1 antigen assay in a region where all dengue serotypes circulate year round and clinical severity could be evaluated 2) Assess the sensitivity of this assay in secondary cases according to pre-illness and acute dengue immune status 3) Determine the sensitivity of this assay according to clinical severity.

Materials and Methods

Results

NS1 sensitivity according to disease severity and gender

As the sensitivity of the NS1 assay was 100% in primary infections, only secondary infections were included in this analysis. The sensitivity among DF cases (25%, 95% CI: 16.7-33.3) was significantly lower compared to DF+HM cases (55%, 95% CI: 33.2-76.8) and DHF I and DHF II cases (50%, CI: 36.9-63.1). Additionally, the sensitivity was higher in patients with platelet counts below 100,000/mm³ at presentation (48% vs. 27.1%, p=0.00) or at time of platelet nadir (50% vs.16.2%, p=0.00). A higher sensitivity was also found in cases with spontaneous hemorrhagic manifestations, hemoconcentration or effusion, but these differences were not significant (Figure 1). Data from a subset of patients enrolled in study one and two in which patients were identified early, revealed that the mean and range interval between the detection of NS1 and thrombocytopenia among 28 patients was 1.8 (0-5) days, whereas the time between the detection of NS1 and evidence of plasma leakage among 17 DHF patients was 3.3 (1-5) days. The sensitivity of the NS1 assay was significantly higher in females than males with values of 49.1% (95% CI: 36.1-62.1) and 29.6% (95% CI: 21.6-37.6) respectively (p=0.01). This gender difference remained present when analysis was restricted to those with manifestations of more severe disease (Figure 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Sensitivity of the NS1 antigen test according to diagnosis category and clinical variables in secondary dengue cases.

https://doi.org/10.1371/journal.pone.0080891.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. The sensitivity of the NS1 antigen test according to gender and clinical variables in secondary infections.

https://doi.org/10.1371/journal.pone.0080891.g002

Discussion

NS1 tests are increasingly being used as a rapid and inexpensive diagnostic tool for acute DF. We found that its sensitivity in a dengue hyper-endemic area (Indonesia) where most infections are secondary was low with an overall sensitivity of 46.8%. In samples from primary infections, the sensitivity was 100% indicating that it is a good diagnostic tool for patients from non-endemic areas, such as in travelers[29]. Our findings are also in line with reports from other countries that found a low sensitivity in secondary infections [4,13,20]. Apart from primary or secondary infection status, our data also suggest that sensitivity was affected by dengue immune status, infecting dengue serotype, gender and severity of illness.

Our data demonstrated poor performance of the NS1 test in acute sera with high titers of HI dengue antibodies. Further analysis showed that NS1 antigen was undetectable in nearly all acute sera from patients with pre-illness neutralizing antibodies to more than one serotype, but was detectable in more than half of those with only one serotype. These findings are in line with the hypothesis that immune-complexes are formed in secondary infections and high levels of antibodies capture and bind soluble NS1 (sNS1) resulting in a reduction of circulating NS1[17]. Therefore, the use of a NS1 assay in areas where post-secondary dengue infection is common [30,31] has limited value. The serotype of a previous dengue infection may also influence the sensitivity of NS1 detection in secondary infections as we found the sensitivity was significantly higher in acute samples from patients with pre-illness neutralizing antibodies to DENV-2 (88.9%) compared to DENV-1 (11.1%). It has been reported that in comparison to DENV-1 and DENV-3, DENV-2 infections result in much lower concentrations of sNS1[32]. Thus, during primary DENV-2 infections, the amount of sNS1 produced may be unable to generate sufficient NS1 antibodies to form immune-complexes with sNS1 in secondary infections. The influence of pre-existing antibodies to other flaviviruses that may circulate in the region such as Japanese encephalitis as a result of natural infection or vaccination may also affect the sensitivity of NS1 and should be considered. The data presented here are consistent with previous studies that report a higher sensitivity of the NS1 test in DENV-3 and DENV-1 infections compared to DENV-2 and DENV-4 infections [12,14]. As the specimens used in this study included all DENV serotypes, the overall sensitivity was not affected by a predominating serotype as has been reported elsewhere [7].

Our study also demonstrated that in secondary cases, the sensitivity is better in patients with more severe illness. The sensitivity of the test in the DF group was significantly lower compared to the other groups (DF+HM or DHF). DSS cases were excluded from this analysis, as the number of DSS cases was too small (n=2). The association between NS1 and clinical severity has been studied before with discordant results[4,6,7,15,21]. The differences might be due to inadequate number of cases[21] or the application of different clinical category guidelines [4]. NS1 sensitivity was also significantly higher in patients with platelet counts below 100,000/mm³ measured at presentation only or nadir platelet counts below 100,000/mm³ if serial measurements were done. A positive NS1 result in secondary infections may therefore be a predictor for low platelet count. Although not significantly different, sensitivity tended to be higher in cases with spontaneous bleeding, effusion and hemoconcentration. The association between NS1 and clinical severity is thought to be related to the cross-reactivity of anti-dengue NS1 antibodies with platelets and endothelial cells, and the role of NS1 in the occurrence of vascular leakage through complement and cytokine responses [33,34]. Recent findings have also shown that NS1 positivity is correlated with higher viremia or antigenemia which was found in more severe cases[13,15].

The overall sensitivity of the NS1 test and its sensitivity in secondary infections in our study is lower than what has previously been published. There are several possible reasons for this difference. This and other studies have demonstrated that the sensitivity of NS1 in secondary infections is much lower than in primary infections. As a consequence, the overall sensitivity will be affected by the proportion of primary or secondary infections in the sample set used. Compared to studies conducted in other endemic countries, our study has a low proportion of primary infections. Our study also included outpatients (36.4%). As has previously been reported and is found in this study, NS1 sensitivity is associated with disease severity. Therefore, outpatients in our study, which were mostly mild cases, may have contributed to the lower NS1 sensitivity. Additionally, we found that among secondary cases, post-secondary (≥2) cases (cases with pre-illness neutralizing antibodies to more than one serotype) had lower sensitivities than secondary cases with pre-illness neutralizing antibodies to only one serotype. Based upon the data from the 47 subjects that were tested for pre-illness neutralizing antibodies and the mean age of our volunteers, we suspect that many of the cases in our study were post-secondary infections. Finally, the proportion of secondary infections caused by DENV-2 and DENV-4, serotypes with lower sensitivities in our study, was higher than in other studies.

The strengths of this study are: 1) work was conducted in a setting where all dengue serotypes were equally distributed and the ratio between primary and secondary infections reflected the regional epidemiology 2) accurate clinical outputs were available as patients were monitored closely 3) this study provides a wealth of laboratory data to assess the correlation between NS1 detection and serotype distribution, and dengue immune status including serotype specific pre-existing immunity, which have not been reported elsewhere. This study has also several limitations. First, male participants were more predominant than females. However, despite fewer samples from females, the detection of NS1 was significantly higher in females than males. Second, ultrasonography was not performed on a subset of volunteers. As a consequence, several DF+HM cases might be more accurately included in the DHF group. However, the numbers are probably small as daily hemoconcentration was performed on all individuals. Third, PRNT was only conducted on pre-illness sera from 47 patients. Although, the results showed a strong association between sensitivity and the presence of neutralizing antibodies to DENV-2, further study is needed.

In conclusion, the NS1 test is not recommended as a standalone diagnostic test for dengue infection in regions where secondary infections predominate. However, as the sensitivity and specificity in acute specimens from primary infections is excellent, use of the test may prove of value to travelers returning from dengue endemic areas. Furthermore, as the sensitivity in secondary cases is associated with clinical severity, NS1 may be useful to predict severe disease. However, further studies with larger sample sizes and in other geographic regions are needed to confirm this association.

Supporting Information

Acknowledgments

We thank the dengue research team at NAMRU-2, Jakarta, at Hasan Sadikin Hospital/Padjadjaran University and Immanuel Hospital and the factory clinic at PT Grandtex Indonesia, Naintex, and bintang Agung. Competing interests: None declared. Ethical approval: These studies were approved by the Institutional Review Board of NAMRU-2 (N2.2006.0001, N2.2004.0010, N2.2004.0008) and at the National Institute of Health Research and Development, Ministry of Health, Indonesia (KS.02.01.2.1.2776, KS.02.01.2.1.3461 and KS.02.01.2.1.2336) in compliance with all U.S. Federal Regulations governing the protection of human subjects. Disclaimer: the views expressed in this article are those of the author and do not necessarily reflect the official policy or position of the Department of the Navy, Department of Defense, nor the U.S. Government. Some of the authors are military service members or U.S. Government employees. This work was prepared as part of their official duties. Title 17 U.S.C. §105 provides that ‘Copyright protection under this title is not available for any work of the United States Government.’ Title 17 U.S.C. §101 defines a US Government work as a work prepared by a military service member or employee of the U.S. Government as part of that person’s official duties.