Diagnostic accuracy of NS1 ELISA and lateral flow rapid tests for dengue sensitivity, specificity and relationship to viraemia and antibody responses

Abstract

Background: Dengue is a public health problem in many countries. Rapid diagnosis of dengue can assist patient triage and management. Detection of the dengue viral protein, NS1, represents a new approach to dengue diagnosis.

Methodology/principal findings: The sensitivity and specificity of the Platelia NS1 ELISA assay and an NS1 lateral flow rapid test (LFRT) were compared against a gold standard reference diagnostic algorithm in 138 Vietnamese children and adults. Overall, the Platelia NS1 ELISA was modestly more sensitive (82%) than the NS1 LFRT (72%) in confirmed dengue cases. Both ELISA and LFRT assays were more sensitive for primary than secondary dengue, and for specimens collected within 3 days of illness onset relative to later time points. The presence of measurable DENV-reactive IgG and to a lesser extent IgM in the test sample was associated with a significantly lower rate of NS1 detection in both assays. NS1 positivity was associated with the underlying viraemia, as NS1-positive samples had a significantly higher viraemia than NS1-negative samples matched for duration of illness. The Platelia and NS1 LFRT were 100% specific, being negative in all febrile patients without evidence of recent dengue, as well as in patients with enteric fever, malaria, Japanese encephalitis and leptospirosis.

Conclusions/significance: Collectively, these data suggest NS1 assays deserve inclusion in the diagnostic evaluation of dengue patients, but with due consideration for the limitations in patients who present late in their illness or have a concomitant humoral immune response.

Figure 1. Relationship between day of illness and NS1 sensitivity.

Shown is percentage sensitivity of (A) Platelia ELISA and (B) LFRT assays by day of illness in patients with confirmed dengue (n = 125).

Figure 2. Viral loads by NS1 status in the Platelia ELISA at the time of study enrolment or after 3 days of illness.

Shown is the mean (interquartile and range) viraemia level in NS1 positive and NS1 negative (Platelia ELISA) patients with a measurable viraemia (n = 111) at (A) the time of study enrolment or (B) after 3 days of illness durations. The limit of detection of the assay is shown with a dashed line. Viraemia levels were significantly higher in NS1 positive patients relative to NS1 negative patients (Mann-Whitney test). The same observations with regard to viraemia levels were made with the NS1 LFRT (data not shown).

Figure 3. NS1 sensitivity of the Platelia ELISA in the context of viral serotype and humoral immune response.

Shown in (A) is the sensitivity of NS1 detection in the enrolment sample according to the infecting serotype identified by real-time RT-PCR (results for DENV-4 not shown as the sample size was small: n = 3). NS1 detection in DENV-2 infected patients was significantly less sensitive than for DENV-1 or DENV-3. The proportion of patients with detectable DENV-reactive (B) IgG or (C) IgM antibodies (measured by capture ELISAs) in the test sample was also related to the infecting serotype. Test samples from DENV-2 infected patients were more likely to have measurable levels of DENV-reactive IgG but not IgM, albeit this was not statistically significant.