Search

Background: Dengue is one of the most common diseases in the tropics and subtropics. Whilst mortality is a rare event when adequate supportive care can be provided, a large number of patients get hospitalised with dengue every year that places a heavy burden on local health systems. A better understanding of the support required at the time of hospitalisation is therefore of critical importance for healthcare planning, especially when resources are limited during major outbreaks. Methods: Here we performed a retrospective analysis of clinical data from over 1500 individuals hospitalised with dengue in Vietnam between 2017 and 2019. Using a broad panel of potential biomarkers, we sought to evaluate robust predictors of prolonged hospitalisation periods. Results: Our analyses revealed a lead-time bias, whereby early admission to hospital correlates with longer hospital stays ‐ irrespective of disease severity. Importantly, taking into account the symptom duration prior to hospitalisation significantly affects observed associations between hospitalisation length and previously reported risk markers of prolonged stays, which themselves showed marked inter-annual variations. Once corrected for symptom duration, age, temperature at admission and elevated neutrophil-to-lymphocyte ratio were found predictive of longer hospitalisation periods. Conclusion: This study demonstrates that the time since dengue symptom onset is one of the most significant predictors for the length of hospital stays, independent of the assigned severity score. Pre-hospital symptom durations need to be accounted for to evaluate clinically relevant biomarkers of dengue hospitalisation trajectories. Author summary: Dengue places a significant burden on healthcare settings. Especially in low and middle income settings and during large outbreaks, allocation of limited resources to those at high risk of morbidity and mortality can be critically important. Various risk factors of severe infection outcomes and hospitalisation, such as secondary heterologous infection, have been described, yet reliable biomarkers predictive of prolonged stays once hospitalised are still lacking. In this work we analysed dengue hospitalisation data collected over a period of three consecutive years in Northern Vietnam, which revealed an unexpected negative correlation between dengue severity and length of hospitalisation. Further analysis showed that this was primarily driven by a longer period between symptom onset and admission in those patients with a higher severity score. Moreover, we found that this delay negated other observed correlates of prolonged hospital stays, which themselves revealed significant inter-annual variations. Taken together, this work demonstrates that time to admission is one of the strongest predictors of hospitalisation length and that this needs to be taken into consideration for finding reliable biomarkers of predicted healthcare needs in patients admitted to hospital due to dengue. [ABSTRACT FROM AUTHOR]