Severe Dengue: Developing a Universally Applicable Simple Prediction Model

Excerpt

Globally dengue causes 50 million infections per year and an estimated 20,000 deaths annually, posing a large disease and economic burden. In 2009, World Health Organization (WHO) proposed a classification categorising the disease as dengue with or without warning signs and severe dengue, to enhance early recognition of severe dengue [1]. The 2009 classification suggests 7 clinical signs as warning signs for severe dengue based on a multicenter study in Southeast Asia and Latin American countries [2] and global expert consensus. WHO guidelines suggest that presence of any warning sign should lead to admission to a healthcare facility and intravenous fluid therapy [1]. A systematic review showed that sensitivity of 2009 classification to identify severe dengue ranged between 59 and 98% (88%/98%: prospective studies), and specificity between 41 and 99% (99%: prospective studies) [3]. Although the WHO 2009 guidelines’ system of early warning signs is easy to use and sensitive in identifying severe Dengue, and beneficial to triage and case management, there remain issues with its applicability. It has resulted in large numbers of admissions of dengue patients to hospitals imposing a huge economical and physical burden, particularly in South East Asia. A lack of a simple tool to discriminate mild from life threatening infection has led to unnecessary hospitalization of dengue patients. Quantitative research into development of predictive tools to triage the patients at greater risk of severe dengue and the predictive value of these warning signs on patient outcomes may optimize usage of limited healthcare resources in resource-poor settings [4, 5]. …