Comparison of two commercially available dengue virus (DENV) NS1 capture enzyme-linked immunosorbent assays using a single clinical sample for diagnosis of acute DENV infection

Abstract

Dengue virus (DENV) nonstructural protein 1 (NS1) has shown promise as a novel diagnostic marker of acute DENV infection. Current techniques used to diagnose acute DENV infection, including virus isolation and reverse transcription-PCR (RT-PCR), are costly and difficult to perform, while traditional serological assays have low sensitivities during the acute stage of infection. Two commercially available NS1 antigen capture enzyme-linked immunosorbent assays (ELISAs), the Platelia dengue NS1Ag test (Bio-Rad Laboratories, Marnes La Coquette, France) and the Pan-E dengue early ELISA test (Panbio Diagnostics, Brisbane, Australia), were evaluated against a well-characterized panel of 208 real-time RT-PCR- and virus isolation-positive sera, as well as 45 real-time RT-PCR- and serologically negative sera from patients with other acute febrile illnesses. The overall sensitivities were 64.9% (95% confidence interval [CI(95)], 58.2 to 71.1%) for the Panbio test and 83.2% (CI(95), 77.5 to 87.7%) for the Bio-Rad test, with interserotype variation, especially for DENV serotype 4. Predictive models were constructed to identify factors that had a significant influence on a test's outcome with respect to this panel of samples in order to identify the conditions in which the test will be most effective as a diagnostic tool. The immunoglobulin G titer was found to be the only covariate that significantly influenced results in the Bio-Rad test, while serotype and the day postonset were found to significantly influence results in the Panbio test. We concluded that the NS1 capture ELISA is a useful tool that can improve testing algorithms to diagnose DENV infection in single samples from acute and early convalescent cases.

FIG. 1.

Predictive model for the Bio-Rad kit based on the covariate IgG titer. IgG titer is represented as an integer value along the x axis, where 0 is <1:40, 1 is 1:40, 2 is 1:160, 3 is 1:640, 4 is 1:2,560, 5 is 1:10,240, 6 is 1:40,960, 7 is 1:163,840, and 8 is ≥1:655,360. The frequency of a positive result is plotted along the y axis. The solid line represents the frequency of a positive result, as predicted by the model, and the dotted line represents the actual data. Generally, as the IgG titer increases, the frequency of a positive result decreases.

FIG. 2.

Predictive model for the Panbio kit based on the covariates DPO and serotype. DPO is plotted along the x axis, with the frequency of a positive result plotted along the y axis. The solid lines represent the frequencies of positive results for DENV-1, DENV-2, DENV-3, and DENV-4 predicted by the model for each DPO. The dotted lines represent the actual data. In general, as DPO increases, so does the frequency of a positive result. The frequency of a positive result was significantly lower for DENV-4 samples than for DENV-1 to -3 samples.

FIG. 3.

Comparison of NS1 detection for each DPO. DPO is plotted along the x axis, with frequency of NS1 detection plotted along the y axis, for the Bio-Rad and Panbio kits.