Dengue Evaluation of Multi-State Models (DENEM Study): A Prospective Observational Study of Patients Hospitalised With Dengue Vascular Leak in Nha Trang, Vietnam, to Develop Novel Statistical Methodology

Dengue is a life-threatening infection caused by a virus, spread between humans by the bite of mosquitoes. It is present throughout the tropics, including in Vietnam, where cases have dramatically increased in recent years. Dengue causes severe illness predominantly through vascular leak, where patients' blood vessels break down, becoming leaky, leading to fluid from the vessels moving into tissues and organs such as the lungs.

However, it is still not fully understand how the virus causes vascular leak, or in which patients it is most likely to occur in. There are no licensed treatments for vascular leak. This is because the mechanism of vascular leak is incompletely understood, making therapeutic targeting difficult. Additionally, when drugs are trialled, many trials have been poorly designed and not included enough patients.

To answer our research questions, the investigators will recruit 142 patients admitted to hospital with dengue in Nha Trang, Vietnam who have dengue vascular leak. If they are happy to enter the study, data will be recorded that is already being collected as part of their hospital admission; this will include clinical data (such as blood pressure, pulse and treatments given) and the results of their blood tests. Investigators will also run tests beyond what they would normally have in hospital; this will include the results of regular ultrasound scans and biomarker blood tests (small molecules that can be detected in their blood in response to stress and vascular leak). Most patients have blood tests daily in hospital, and clinicians will aim to take the extra tubes of blood required at the same time, to minimise the number of extra procedures requested from participants.

Investigators will put this data it into a statistical model of dengue vascular leak they have been developing, called a multi-state model. These models have previously been used in other areas of medicine, but this would be their first application in dengue and infectious diseases research. Multi-state models aim to track how patients move through different stages of illness over time. Models such as this make better use of all the information collected during a patient's illness. Rather than only looking at a single outcome, such as whether a patient had recovered by a certain day, or how long recovery took, they track how patients move through different stages of dengue over time and how long they spend in each stage. This may give a more complete picture of how treatments affect the course of illness. If successful, it could improve how future dengue clinical trials are designed, helping researchers detect whether new treatments work more quickly and with fewer volunteers.

Investigators will also use data and samples collected from this study to explore why vascular leak occurs and evaluate new ways of diagnosing dengue, particularly in low-resource settings. To achieve this, investigators will need to compare patients who have dengue to people who have not got dengue, looking for differences. As such 93 people who do not have dengue (called "control" participants) will be recruited. Some will have a fever from another cause, and others will be healthy volunteers. Comparing these control populations with patients who have dengue helps us understand which findings are specific to dengue and how accurate new dengue tests are. People in these groups will only have a single small blood sample taken and will not receive any treatment or need follow-up visits.