Djossou F, Vesin G, Elenga N, Demar M, Epelboin L, Walter G, Abboud P, Le-Guen T, Rousset D, Moreau B, Mahamat A, Malvy D, and Nacher M
Background: Identifying patients at risk of developing severe dengue is challenging. The objective of the present study was to determine the incidence of hypotension and its predictive factors during the Dengue 2 epidemic in 2013., Methods: In 2013, a longitudinal study was performed using data from all confirmed cases of dengue seen in Cayenne General Hospital. The analysis used Cox proportional modeling to obtain adjusted hazards ratios for hypotension., Results: A total of 806 confirmed patients were included 78 (9.6%) of whom developed hypotension. Extensive purpura, cutaneomucous hemorrhage, serous effusion and age 1-15 years were associated with subsequent hypotension whereas 'aches' and a rash were associated with a lower incidence of hypotension. The biological variables independently associated with hypotension were: increase of hematocrit, low protein concentrations, low sodium concentration and lymphocytes over 1400/ml. A risk score was computed from the scaled Cox model coefficient., Conclusions: From a clinician's perspective, extensive purpura, cutaneomucous hemorrhage, serous effusion, age 1-15 years, hematocrit increase, low protein, low sodium, lymphocytosis and the absence of aches or of a rash, may be important warning signs to predict subsequent hypotension and shock. Over half of the patients with the highest risk score subsequently developed hypotension. The prognostic score had a 48.2% sensitivity with less than 10% of false positives. This score requires external validation before its impact on clinical practice is evaluated., (© The Author 2017. Published by Oxford University Press on behalf of Royal Society of Tropical Medicine and Hygiene. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.)