Dengue is a rapidly expanding arboviral illness with major public-health impact, marked by a small proportion of patients progressing to severe disease characterized by profound plasma leakage and shock (severe dengue) (WHO 2025) (1). In modern classifications and clinical pathways, severe dengue is defined by severe plasma leakage leading to shock or respiratory distress, severe bleeding, or severe organ involvement (2). The pathobiology of the capillary leak syndrome (CLS) in dengue is driven by transient endothelial dysfunction with heightened permeability mediated by cytokines, the dengue NS1 antigen, complement activation, and lipid mediators such as platelet-activating factor (Tayal 2022). Critically, ultrasound studies show that plasma leakage often begins around defervescence and may be detectable before hematocrit changes, providing a window for early risk stratification (4). Conventional sonographic markers of leakage—pleural effusion, ascites, and gallbladder-wall thickening (GBWT)—are well described and correlate with severity (5). However, these are largely structural or binary findings that may appear late or vary with operator and hydration status, and they do not directly quantify the hemodynamic burden of venous congestion during the critical phase (6). Portal venous Doppler offers a physiologic window into splanchnic hemodynamics that might map more closely to the dynamics of dengue-associated CLS. The normal portal vein waveform is minimally pulsatile and hepatopetal; with rising right-atrial pressure and systemic venous congestion, portal pulsatility increases and may even become biphasic or to-and-fro (7). Quantitative indices include the portal vein pulsatility index (PVPI = [Vmax–Vmin]/Vmax) and the portal vein congestion index (PVCI = cross-sectional area/mean velocity), both of which rise with venous congestion and reduced forward flow (8). Outside dengue, PVPI and related venous excess ultrasound (VExUS) components track fluid responsiveness, right-heart loading, and congestion in perioperative and critical-care cohorts (9). Within dengue specifically, ultrasound has been extensively studied for GBWT and third-spacing, but far fewer studies have interrogated portal venous Doppler hemodynamics as early predictors of CLS. A recent single-center study suggested that a higher PVCI and GBWT discriminated patients who developed CLS, with congestion index showing strong area-under-curve performance on ROC analysis; nonetheless, thresholds varied and external validation was lacking (IOSR JDMS 2025). Broader evidence syntheses also emphasize that “plasma leakage” cohorts are inconsistently defined across studies, which complicates biomarker and imaging validation (6). This gap matters clinically. Dengue fluid management balances avoidance of hypovolemia against the risks of fluid overload as capillary integrity recovers; tools that reflect real-time venous congestion could refine titration and reduce iatrogenic harm (3). While GBWT and pleural/abdominal free fluid signal established leakage, portal Doppler may provide earlier, quantitative cues of splanchnic venous hypertension before overt third spacing. Yet, in dengue, normative ranges, reproducible cut-offs, and the temporal behavior of PVPI and PVCI across febrile, critical, and recovery phases remain poorly characterized, and correlations with clinical end-points (shock, need for colloids, ICU admission), laboratory markers (hematocrit rise, albumin fall), and other ultrasound indices (IVC collapsibility, VExUS grade) are inconsistent or unreported (10). There is also limited comparison of portal Doppler against commonly used sonographic surrogates (GBWT, pleural effusion) for predicting CLS on multivariable analyses, and few prospective studies establish how dynamic changes in portal indices track the transition into and out of the critical phase. Aim of the present study was to characterize hemodynamic alterations in portal venous Doppler among laboratory-confirmed dengue patients across clinical phases and to determine their correlation with capillary leak syndrome, testing whether quantitative portal indices (PVPI and PVCI) independently predict CLS compared with conventional sonographic markers (GBWT, pleural effusion/ascites) and bedside measures of intravascular volume (IVC collapsibility), while defining pragmatic cut-offs for early risk stratification and fluid-management guidance.

Study Design and Setting: This was a prospective observational study conducted in the Department of Radiology. The study was approved by the Institutional Ethics Committee, and informed consent was obtained from all participants or their legal guardians prior to enrollment. Study Population and Sample Size: A total of 75 patients with laboratory-confirmed dengue infection (positive NS1 antigen and/or IgM ELISA) were included in the study. Patients were enrolled if they presented within the first week of illness and underwent abdominal ultrasound and Doppler evaluation. Inclusion Criteria: • Laboratory-confirmed dengue infection (NS1 antigen and/or IgM ELISA positive). • Age ≥ 18 years. • Patients presenting within the first 7 days of illness. • Patients willing to undergo ultrasound and Doppler evaluation with informed consent. Exclusion Criteria: • Known chronic liver disease or established portal hypertension. • Right-sided heart failure or any condition causing systemic venous congestion (e.g., tricuspid regurgitation). • Chronic kidney disease or baseline renal dysfunction unrelated to dengue. • Previous abdominal surgery significantly altering portal or IVC anatomy. • Uncooperative patients or technically inadequate Doppler studies. Clinical and Laboratory Assessment: Baseline demographic data, clinical presentation, and routine laboratory investigations (hematocrit, platelet count, liver function tests, renal function tests) were documented. Capillary leak syndrome (CLS) was defined according to WHO criteria as a hematocrit rise >20% from baseline and/or presence of serosal effusion with clinical features such as hypotension, shock, or evidence of plasma leakage on imaging. Ultrasound and Doppler Protocol: All patients underwent abdominal ultrasonography and portal venous Doppler using a curvilinear transducer (3–5 MHz). Standardized scanning was performed in the supine position after a fasting period of at least 4 hours. Portal vein diameter, mean velocity, and spectral waveform patterns were recorded. The Portal Vein Pulsatility Index (PVPI) was calculated as (Vmax – Vmin)/Vmax, and the Portal Vein Congestion Index (PVCI) was calculated as the cross-sectional area divided by mean velocity. Additional markers such as gallbladder wall thickness (GBWT), pleural effusion, and ascites were documented. Inferior vena cava (IVC) diameter and collapsibility index were also assessed to correlate intravascular volume status. Statistical Analysis: Patients were stratified into two groups based on the presence or absence of CLS. Continuous variables were expressed as mean ± SD and compared using Student’s t-test or Mann–Whitney U test, as appropriate. Categorical variables were compared using Chi-square or Fisher’s exact test. Receiver operating characteristic (ROC) curves were constructed for PVPI and PVCI to assess their diagnostic performance in predicting CLS. A p value <0.05 was considered statistically significant.

Table 1. Laboratory Parameters in Dengue Patients with and Without Capillary Leak Syndrome (CLS) Parameter Dengue with CLS (n = 35) Dengue without CLS (n = 40) Hematocrit (%) 44.8 ± 5.2 38.6 ± 4.3 Platelet count (×10⁹/L) 52 ± 24 104 ± 36 Serum AST (U/L) 182 ± 95 74 ± 38 Serum ALT (U/L) 134 ± 70 62 ± 29 Serum creatinine (mg/dL) 1.24 ± 0.36 0.92 ± 0.22 Serum albumin (g/dL) 2.9 ± 0.5 3.7 ± 0.6 Patients who developed CLS demonstrated a significantly higher hematocrit (44.8 ± 5.2% vs 38.6 ± 4.3%), reflecting hemoconcentration due to plasma leakage. Platelet counts were markedly lower in the CLS group (52 ± 24 ×10⁹/L vs 104 ± 36 ×10⁹/L), consistent with severe dengue-associated thrombocytopenia. Liver enzymes were elevated in CLS, with serum AST (182 ± 95 U/L) and ALT (134 ± 70 U/L) nearly double those in non-CLS patients, indicating hepatic involvement. Renal dysfunction was also more frequent, as evidenced by higher serum creatinine (1.24 ± 0.36 mg/dL vs 0.92 ± 0.22 mg/dL). Importantly, serum albumin levels were lower in CLS patients (2.9 ± 0.5 g/dL vs 3.7 ± 0.6 g/dL), supporting the presence of vascular leakage and protein extravasation (Table 1). Table 2. Portal Venous Doppler Parameters in Dengue Patients Parameter Dengue with CLS (n ≈ 35) Dengue without CLS (n ≈ 40) Portal vein diameter (mm) 12.4 ± 1.8 10.8 ± 1.6 Mean portal vein velocity (cm/s) 11.2 ± 2.9 16.7 ± 3.8 Spectral waveform pattern Increased pulsatility; biphasic in some cases Predominantly monophasic hepatopetal PVPI 0.42 ± 0.09 0.23 ± 0.07 PVCI (cm²/cm/s) 0.21 ± 0.06 0.13 ± 0.04 Table 2 shows, patients with CLS had a larger portal vein diameter (12.4 ± 1.8 mm vs 10.8 ± 1.6 mm) and significantly reduced mean portal vein velocity (11.2 ± 2.9 cm/s vs 16.7 ± 3.8 cm/s), reflecting venous stasis and congestion. The spectral waveform pattern showed increased pulsatility in CLS patients, with some demonstrating biphasic flow, in contrast to the predominantly monophasic hepatopetal waveform in non-CLS patients. Quantitative indices further highlighted hemodynamic differences: the portal vein pulsatility index (PVPI) was nearly double in CLS patients (0.42 ± 0.09 vs 0.23 ± 0.07), while the portal vein congestion index (PVCI) was significantly elevated (0.21 ± 0.06 vs 0.13 ± 0.04). Table 3. Sonographic Markers of Plasma Leakage and Volume Status in Dengue Patients Parameter Dengue with CLS (n = 35) Dengue without CLS (n = 40) Gallbladder wall thickness (mm) 6.8 ± 1.5 3.1 ± 0.9 Pleural effusion (%) 71% 14% Ascites (%) 63% 20% IVC diameter (mm) 18.0 ± 2.7 14.5 ± 2.2 IVC collapsibility index (%) 19 ± 7 34 ± 10 Gallbladder wall thickness was markedly higher in the CLS group (6.8 ± 1.5 mm vs 3.1 ± 0.9 mm), supporting its role as an early and reliable marker of plasma leakage. Pleural effusion (71% vs 14%) and ascites (63% vs 20%) were significantly more frequent in CLS patients, reflecting third-space fluid accumulation during the critical phase. Inferior vena cava (IVC) measurements also demonstrated important differences: CLS patients had a larger IVC diameter (18.0 ± 2.7 mm vs 14.5 ± 2.2 mm) and a lower collapsibility index (19 ± 7% vs 34 ± 10%), consistent with venous congestion and reduced volume reserve (Table 3). Table 4. Comparison of Key Parameters between Dengue Patients With and Without CLS Parameter CLS (n=35) Non-CLS (n=40) p Value Hematocrit (%) 44.8 ± 5.2 38.6 ± 4.3 <0.001 Platelet count (×10⁹/L) 52 ± 24 104 ± 36 <0.001 Serum AST (U/L) 182 ± 95 74 ± 38 <0.001 Serum ALT (U/L) 134 ± 70 62 ± 29 <0.01 Serum creatinine (mg/dL) 1.24 ± 0.36 0.92 ± 0.22 <0.01 PVPI 0.42 ± 0.09 0.23 ± 0.07 <0.001 PVCI (cm²/cm/s) 0.21 ± 0.06 0.13 ± 0.04 <0.001 GBWT (mm) 6.8 ± 1.5 3.1 ± 0.9 <0.001 Pleural effusion (%) 71% 14% <0.001 Ascites (%) 63% 20% <0.001 IVC diameter (mm) 18.0 ± 2.7 14.5 ± 2.2 <0.001 IVC collapsibility index (%) 19 ± 7 34 ± 10 <0.001 Patients with CLS showed significantly higher hematocrit (44.8 ± 5.2% vs 38.6 ± 4.3%, p <0.001) and markedly lower platelet counts (52 ± 24 vs 104 ± 36 ×10⁹/L, p <0.001), reflecting hemoconcentration and thrombocytopenia. Liver involvement was more pronounced in CLS, with serum AST and ALT nearly doubled compared to non-CLS patients. Renal impairment was evident through elevated creatinine (1.24 ± 0.36 vs 0.92 ± 0.22 mg/dL, p <0.01). Portal venous Doppler revealed striking hemodynamic alterations: PVPI and PVCI were significantly elevated in CLS patients, confirming venous congestion and altered flow dynamics. Conventional sonographic markers also differed: gallbladder wall thickness was more than doubled (6.8 ± 1.5 vs 3.1 ± 0.9 mm, p <0.001), while pleural effusion (71% vs 14%) and ascites (63% vs 20%) were far more prevalent in the CLS group. IVC assessment supported these findings, with CLS patients showing increased diameter and markedly reduced collapsibility, indicating impaired intravascular volume regulation. Table 5. Correlation Analysis of Doppler Indices with Hematocrit, Albumin, and IVC Collapsibility (CLS group, n=35) Correlation Pair r (Correlation Coefficient) p value PVPI vs Hematocrit –0.12 0.496 PVPI vs Albumin +0.02 0.901 PVPI vs IVC Collapsibility 0.00 0.995 PVCI vs Hematocrit –0.29 0.091 PVCI vs Albumin –0.18 0.292 PVCI vs IVC Collapsibility +0.21 0.224 NS = not statistically significant (p > 0.05). Neither PVPI nor PVCI demonstrated statistically significant correlations with hematocrit, serum albumin, or IVC collapsibility (all p >0.05). PVPI showed a weak negative correlation with hematocrit (r = –0.12, p = 0.496) and a negligible association with albumin (r = +0.02, p = 0.901) and IVC collapsibility (r ≈ 0.00, p = 0.995). Similarly, PVCI demonstrated weak negative correlations with hematocrit (r = –0.29, p = 0.091) and albumin (r = –0.18, p = 0.292), and a weak positive correlation with IVC collapsibility (r = +0.21, p = 0.224). Although these trends suggest some physiological associations, none reached statistical significance (Table 5).

In this prospective cohort of 75 laboratory-confirmed dengue patients studied in the radiology department, we evaluated splanchnic venous hemodynamics using portal venous Doppler alongside conventional sonographic markers of plasma leakage and bedside indices of volume status. The principal finding was the strong diagnostic performance of PVPI and PVCI for predicting capillary leak syndrome (CLS): PVPI showed an AUC of 0.962 with an optimal cutoff of 0.324 (sensitivity 85.7%, specificity 97.5%), and PVCI showed an AUC of 0.911 with a cutoff of 0.152 (sensitivity 88.6%, specificity 82.5%). These results are directionally consistent with the pathophysiologic expectation that venous congestion increases portal waveform pulsatility and reduces forward flow, captured numerically by PVPI and the congestion index, respectively. Prior work in perioperative/critical-care populations has shown that portal vein pulsatility rises with right-heart pressure and venous congestion, supporting its physiologic validity as a dynamic marker (11). Our findings extend the dengue ultrasound literature in two ways. First, most dengue imaging studies have emphasized structural leakage markers pleural effusions, ascites, and gallbladder wall thickening (GBWT) which are reproducibly associated with severity (Dewan et al., 2021; Shah et al., 2018). In the largest scoping review, ascites (≈60%), pleural effusion (≈58%), and GBWT (≈55%) were the most frequent sonographic abnormalities, though the authors underscored heterogeneity in scanning protocols and timing around defervescence that limits pooled estimates (12). In our cohort, we documented these markers and also quantified venous congestion with PVPI/PVCI. Second, only a few dengue-specific studies have tested portal Doppler metrics as predictors of CLS. A recent single-center series reported that the congestion index (PVCI) achieved an AUC ≈0.83 and GBWT ≈0.79 for predicting CLS, concluding that adding portal Doppler improves discrimination beyond structural ultrasound alone (13). Our AUCs for PVPI (0.962) and PVCI (0.911) are higher; likely contributors include standardized fasting measurements, phase-specific scanning during the critical window, and different case-mix. External validation is warranted. Correlation analyses in our dataset showed weak/non-significant linear associations between PVPI/PVCI and hematocrit, albumin, and IVC collapsibility. While this contrasts with the intuitive expectation that greater congestion (higher PVPI/PVCI) accompanies hemoconcentration and lower albumin, there are plausible explanations: (i) temporal dissociation plasma leakage often begins around defervescence and imaging/labs may not be perfectly synchronized; (ii) bimodal physiology patients can transition rapidly from intravascular depletion to fluid-overload as capillary integrity recovers; (iii) mixed determinants albumin is influenced by hepatic synthesis/dilution, while IVC behavior reflects both intrathoracic pressure and venous tone. The dengue pathobiology endothelial dysfunction driven by immune mediators and NS1-complement interactions also predicts phase-dependent, regional heterogeneity in leakage (14), which may blunt cross-sectional correlations with single-timepoint labs. Our results reinforce the value of a hemodynamic ultrasound bundle in dengue: conventional leakage markers (GBWT, effusions/ascites) confirm third-spacing, while portal Doppler provides an earlier, quantitative lens on venous congestion that is less binary and potentially more actionable for fluid titration. The congestion index defined as cross-sectional area divided by mean velocity has a long pedigree (15) and performed well here, aligning with the IOSR study’s ROC data. Meanwhile, PVPI leverages spectral waveform morphology, which becomes more pulsatile or even to-and-fro with raised right-atrial pressure and splanchnic venous hypertension (11). Clinical implications. In settings with high dengue burden, ultrasound is recommended to identify warning signs and guide fluids (WHO, 2025). Integrating PVPI/PVCI thresholds (≈0.32 and ≈0.15, respectively, from our analysis) with GBWT and effusion assessment could help stratify risk at bedside, particularly when hematocrit/albumin are inconclusive or delayed (17). This aligns with WHO guidance emphasizing early recognition of plasma leakage and careful fluid stewardship to avoid both hypovolemia and iatrogenic overload. Limitations. Single-center design, modest sample size, and reliance on single-timepoint measurements limit causal inference. We did not perform right-heart echocardiography or direct central venous pressure measurements that could further anchor PVPI physiology. Inter-observer variability for Doppler sampling (angle correction, sampling site) and GBWT was not formally analyzed. Finally, while our ROC estimates are robust within the cohort, cutoffs require external validation before protocolization.

In dengue, portal venous Doppler adds physiologic insight beyond structural ultrasound. PVPI and PVCI demonstrated high diagnostic accuracy for CLS in our cohort (AUC 0.962 and 0.911, respectively), outperforming traditional single markers and complementing GBWT/effusion assessments. Although cross-sectional correlations with hematocrit, albumin, and IVC collapsibility were weak, the overall signal supports a two-tiered ultrasound approach structural leakage detection plus venous congestion quantification to improve early risk stratification and guide fluid therapy. We recommend prospective, multicenter validation of standardized PVPI/PVCI cutoffs, phase-specific scanning protocols, and integration with clinical/laboratory warning signs to build pragmatic care pathways in resource-constrained settings.

1. WHO. Dengue and severe dengue—fact sheet. Updated Aug 21, 2025. World Health Organization 2. Karyanti MR, Uiterwaal CS, Hadinegoro SR, Widyahening IS, Saldi SR, Heesterbeek JH, Hoes AW, Bruijning-Verhagen P. The value of warning signs from the WHO 2009 dengue classification in detecting severe dengue in children. The Pediatric infectious disease journal. 2024 Jul 1;43(7):630-4. 3. Tayal A, Kabra SK, Lodha R. Management of dengue: an updated review. Indian journal of pediatrics. 2023 Feb;90(2):168-77. 4. Khoo SM, Lee CY. Review of the role of ultrasound in assessing plasma leakage in dengue fever and a proposed scanning protocol. Hong Kong Journal of Emergency Medicine. 2024 Dec;31(6):465-71. 5. Shahsavand Davoudi A, Harandi H, Samiee R, Forghani S, Mohammadi K, Shafaati M. Ultrasound evaluation of gallbladder wall thickness for predicting severe dengue: a systematic review and meta-analysis. The Ultrasound Journal. 2025 Feb 3;17(1):12. 6. Rodrigo C, Sigera C, Fernando D, Rajapakse S. Plasma leakage in dengue: a systematic review of prospective observational studies. BMC infectious diseases. 2021 Oct 20;21(1):1082. 7. Dimopoulos S, Antonopoulos M. Portal vein pulsatility: An important sonographic tool assessment of systemic congestion for critical ill patients. World Journal of Cardiology. 2024 May 26;16(5):221. 8. Ogholoh OD, Zaccheus AO, Ikubor JE, Nwafor NN, Idowu BM, Ogholoh GI. Portal vein congestion index in patients with and without chronic liver disease. Journal of Ultrasonography. 2023 May 11;23(93):53-60. 9. Balan C, Morosanu B, Fodoroiu A, Dobre V, Dumitrache A, Barbulescu RT, Valeanu L, Robu C, Boros C, Nica A, Wong A. Decoding portal vein pulsatility: hemodynamic determinants in a post-hoc analysis of a prospective observational trial. Annals of Intensive Care. 2025 Jun 14;15:81. 10. Chatterjee R, Halder P, Samajdar SS, Ghosal P, Mukherjee S, Jana K, Saha B. Correlation between hematocrit and inferior vena cava collapsibility in adult patients with dengue fever: a prospective, observational study at a tertiary care setup in Eastern India. MRIMS Journal of Health Sciences. 2024 Apr 1;12(2):108-11. 11. Huette P, Guinot PG, Haye G, Moussa MD, Beyls C, Guilbart M, Martineau L, Dupont H, Mahjoub Y, Abou-Arab O. Portal vein pulsatility as a dynamic marker of venous congestion following cardiac surgery: an interventional study using positive end-expiratory pressure. Journal of Clinical Medicine. 2021 Dec 12;10(24):5810. 12. Dewan N, Zuluaga D, Osorio L, Krienke ME, Bakker C, Kirsch J. Ultrasound in dengue: a scoping review. The American journal of tropical medicine and hygiene. 2021 Jan 18;104(3):826. 13. Subramanian A. Role of Portal Venous Doppler in Predicting Capillary Leak Syndrome in Dengue Fever Patients: A Prospective Study (Doctoral dissertation, Kilpauk Medical College, Chennai). 14. Simmons CP, Farrar JJ, Nguyen vV, Wills B. Dengue. N Engl J Med. 2012;366(15):1423-1432. doi:10.1056/NEJMra1110265 15. Moriyasu F, Nishida O, Ban N, Nakamura T, Sakai M, Miyake T, Uchino H. " Congestion index" of the portal vein. American Journal of Roentgenology. 1986 Apr 1;146(4):735-9. 16. Edrees WH, Al-Shehari WA, Al-Haddad AM, Alrahabi LM, Al-Haddad OS, Al-Halani AA. Dengue fever in Yemen: a five-year review, 2020–2024. BMC Infectious Diseases. 2025 Jan 6;25(1):28.