Kinetics of changes of LTE4 throughout the course of illness
LTE4 by itself has shown to cause endothelial dysfunction leading to vascular leakage(16) . Since our results showed that urinary LTE4 levels were significantly higher in patients with DHF, during early illness, we sought to investigate if LTE4 levels were highest during the febrile phase or the critical phase to further understand the possible contribution of LTE4 to disease pathogenesis. In order to study the kinetics of LTE4, we measured the urinary LTE4 levels in 25 patients with DF and 23 patients who developed DHF throughout the course of illness with urine samples taken at 9am in the morning. We found that the urinary LTE4 levels rose in patients with DHF from day 3 to 5 and remained high whereas there was no such rise from day 3 in patients with DF (Fig 4a ). The patients who developed vascular leak (those who developed DHF),entered the critical phase between day 4 and 5 of illness and the rise of LTE4 levels in these patients was when they entered the critical phase. Although the urinary LTE4 levels was high throughout the illness in patients with DHF compared to those with DF, this was not significant at any time point, probably due to the lower number of patients included in the serial sampling, which was likely not sufficiently powered to see changes of statistical significance.
We previously showed that levels of PAF showed a diurnal variation in two cohorts of patients with acute dengue(9, 13). Patients with DHF, had high PAF levels in the morning, which markedly reduced in the afternoon2. Therefore, in order to find out if similar changes were observed with LTE4, levels were measured in the morning (9am) and late afternoon (3pm), in 14 patients with DF and 8 patients with DHF. We found that in some patients while LTE4 levels fell towards the afternoon, in others the LTE4 levels were higher at 3pm (Fig 4b) . Such differences in urinary LTE4 levels were seen in both patients with DF and DHF (Fig 4b). Urinary LTE4 levels were higher in the afternoon samples in 71% (10/14) patients with DF and 50% (4/8) patients with DHF, when compared to the morning urinary LTE4 levels, but overall the timing differences were not significant.
In some patients, several fold changes in the viral loads were also seen between morning and afternoon samples in patients with both DF and DHF(Fig 4c) . Viral loads were higher in the afternoon samples in 38% (5/13) patients with DF and 57% (4/7) patients with DHF, when compared to the morning viral loads. However, these changes in viral loads did not reflect in the changes in urinary LTE4 levels. For instance, while the LTE4 level increased the afternoon samples in 14 patients, the viral loads only increased in 9 samples (Fig 4c ).