PREDICTION OF CAPILLARY LEAKAGE IN PATIENTS WITH DENGUE VIRUS INFECTION: WHAT ELSE BESIDES HEMATOCRIT AND PLATELET COUNTS?

Abstract

INTRODUCTION: Besides clinical examination, hematocrit and platelet counts are often used to predict if a patient with suspected dengue virus infection had dengue hemorrhagic fever. OBJECTIVE: In this study, we investigated the role of Doppler study of the portal vein as a predictor for capillary leakage in these patients. METHODS: Doppler studies of the right portal vein blood flow velocity were performed for 61 patients (aged 10.2 ± 2.9 years; 34 boys and 27 girls) with serologically confirmed dengue virus infection at defervescence. Presence of right pleural effusion was detected by ultrasound in 32 patients 24 to 48 hours later. Binary logistic regression analysis and receiver operating characteristic (ROC) curves were constructed for the following variables as predictors of pleural effusion 24 to 48 hours after defervescence: age, gender, maximum hematocrit level, lowest platelet count, and the velocity of blood flow in the right portal vein (PVDPL). RESULTS: Hematocrit level and PVDPL were independent predictors of pleural effusion. The area under the ROC curve, sensitivity, and specificity for these variables as predictors for right pleural fluid 24 to 48 hours after defervescence are shown in Table 1. CONCLUSIONS: Doppler-derived portal venous blood flow velocity may be used to predict the clinical progression of patients with dengue virus infection. TABLE 1.Variables as Predictors of Pleural EffusionVariablesArea Under ROC CurveCutoffSensitivity, %Specificity, %Hematocrit, %0.79>437283PVDPL, cm/s0.88<15.37279Hematocrit/PVDPL, s/cm0.93>2.668183