Minimisation study of dengue prognostic biomarker panel test

Abstract

Attached

Background: WHO has identified Dengue as the fastest spreading mosquito-borne disease in the world. Approximately 96 million people develop clinical Dengue annually, of which 1 in every 200 proceeds to develop potentially fatal Dengue haemorrhagic fever and/or shock syndrome (DHF/DSS). Depending on the availability of appropriate supportive treatment, the case-fatality rate varies from 3.5% to 50%. The debilitating and painful nature of the disease, together with the lack of an accurate method to predict DHF/DSS, resulted in unnecessary over-hospitalisation. The situation is worsened during epidemic years, in both developed (e.g. Singapore in 2004/2005) and developing countries (e.g. India in 2015) alike. Elective surgeries and non-emergency admissions had to be cancelled to release health resources for Dengue inpatients. Our laboratory previously discovered a panel of novel serum biomarkers that can predict DHF/DSS with a sensitivity and specificity of 90% and 91%, respectively. As majority of Dengue-endemic countries are developing countries, we sought to evaluate how test performance would be affected by using only one biomarker (BM1), so as to make the test more affordable. Methods & Materials: 109 Dengue patients were enrolled from the Colombo South Teaching Hospital and National Hospital of Sri Lanka. The study was approved by the Ethics Review Committee of the University of Sri Jayawardanapura, and all patients provided informed consent. Blood samples were collected when subjects first presented themselves at the hospitals. Serum BM1 concentration was measured using a quantitative ELISA developed in-house. All statistical analyses were performed using R version 3.1.2. Results: 60 subjects were diagnosed with DHF while 49 had classical Dengue fever. Based on only serum BM1 concentration, the performance of the test to predict DHF/DSS was as follows: sensitivity – 73.3%, specificity – 77.6%, positive predictive value (PPV) – 80.0%, and negative predictive value (NPV) – 70.4%. Conclusion: While performance of the BM1 prognostic test (on Sri Lankan patients) could not match that of the panel prognostic test (on Singapore patients), there were considerable improvements in specificity and PPV over current clinical practices used (in Singapore) to determine which Dengue patients to hospitalise (specificity: ≤55%; PPV: ≤34%).