Observational and prospective study carried out at the Internal Medicine Ward of a tertiary hospital between 2022 and 2024. Inclusion criteria were: 1) Age>18 age years old. 2) N-terminal pro-brain Natriuretic Peptide (NT-proBNP)>1000pg/mL. 3) Signs and symptoms of congestion due to HF. 4) Informed consent signed. Exclusion criteria were: 1) Intensive care previous admission. 2) Impaired cognitive or functional status. 3) End-stage kidney disease (dialysis/renal transplant). 4) Advanced Chronic obstructive pulmonary disease, defined as forced expiratory volume in one second (FEV1) <30%.

Assessment of congestion was performed through different methods during the first 24h of admission. Clinical congestion was calculated according to a simplified clinical congestion score3 (Supplementary material, Table S1). Tissular lung congestion was quantified by the identification of significant B lines (≥3 lines) in each quadrant explored (8 in total) and pleural effusion presence. Vascular congestion was estimated by the analysis of the IVC morphology, and both inspiratory and expiratory diameter were calculated through the “M” mode. VExUS protocol was calculated according to Beaubin-Souigny et al.7 Patients were classified on admission as “low vascular congestion” if the total VExUS score was 0-1 and “high vascular congestion” if the VExUS score grade was 2 and 3.

All ultrasound measurements were performed with a “Mylab Omega” ultrasound using sectorial and curvilinear probe (Esaote©) by two researchers with large experience in multimodal assessment of congestion in HF. In the case of low-quality images or not typical venous flow patterns, both physicians discussed the case. In case of discordance between both researchers, the explorations were censured.

Congestive nephropathy was defined as an estimated Glomerular Filtration Rate by CKD-EPI-Creatinine (eGFR)<60mL/min/1.73m2 and/or microalbuminuria>30mg/gr in the presence of a non-continuous renal venous pattern (grades II and III).

A complete blood test analysis was performed on the first morning after admission. NT-proBNP and CA125 were determined (Roche Elecsys® NT-proBNP and ®CA 125 assay). Urinary biomarkers were determined from urinary spot samples collected between two and three hours after the first bolus of endovenous morning furosemide, on the day of admission.

Continuous variables were expressed as the mean±standard deviation or median [Interquartile range] accordingly to the Saphiro-Wilk normality test. Qualitative variables were expressed as percentages.

Baseline patient characteristics were first stratified based on VExUS grading and second, based on the presence of CN. The distinction between groups was evaluated by appropriately using the one-way ANOVA or the Kruskal-Wallis H test for continuous variables and the chi-squared test for categorical variables.

The primary endpoint was the composite of all-cause mortality and/or HF readmissions at 365 days. The difference between groups for the primary outcome was studied with Kaplan-Meier curves and log-rank tests. Potential predictors for the primary end-point were identified through univariate and multivariate logistical regression analyses. List of candidate variables for multivariate analysis, were considered with a p-value ≤0.2 in descriptive analysis between the deceased and surviving groups at 365 days, and variables for which prior biological plausibility had been described.

Estimates are presented with 95% confidence intervals. p-Value<0.05 was considered statistically significant. All analyses were carried out using statistical packages of Python programming language. The study was carried out in compliance with the recommendations contained in the international declaration of Helsinki and approved by the Aragon Health Research Ethics Committee.

According to VExUS protocol on admission, 51% of the patients had low vascular congestion (VExUS 0-1) and 49% had high vascular congestion (VExUS 2-3). Patients with high vascular on admission had a higher prevalence of chronic HF (71.4% vs 45.1%; p=0.008), a higher prevalence of hypertension (93.9% vs 80.4%; p=0.045) and dyslipidemia (67.3% vs 47.1%; p=0.041), and had pacemakers in a higher proportion (22.4% vs 5.9%; p=0.017). According to HF treatment, patients with high vascular congestion were treated with higher proportion of oral loop diuretics (83.7% vs 62.7%; p=0.018), sodium-glucose cotransporter-2 inhibitors (SGLT2i) (42.9% vs 17.4%; p=0.009), angiotensin receptor/neprylisin inhibitor (12.2% vs 2.0%; p=0.044) and non-significant higher prescription of antagonist mineralocorticoid receptor blockers (26.5% vs 11.8%; p=0.060) (Table 1).

Patients with high vascular congestion, showed a higher rate clinical congestion score (5 points [2] vs 4 points [3]; p=<0.001) and higher concentrations of cancer antigen 125 (CA125) (53 [73] U/mL vs 32 [75] U/mL; p=0.045). No differences were identified according to renal function (50.2 [25] mL/min/1.73m2 vs 52.5 [28] mL/min/1.73; p=0.611) or natriuresis (92 [53] mmol/L vs 93 [40] mmol/L; p=0.736), but higher rates of microalbuminuria/creatinine ratio were detected in those patients with higher VExUS scores (130 [230] mg/g vs 80 [99] mg/g; p=0.027). Finally, high vascular congestion group patients needed higher significant doses of endovenous furosemide during the first 72h after admission (545 [895] mg vs 240 [346] mg; p=0.001) (Table 1 & Fig. 1).

Fig. 1.

Bar chart and violin plot regarding variables based on VExUS degree.

A total of 95 patients were finally included to evaluate the presence of CN. A prevalence for CN of 38% was detected in our population on admission. Baseline characteristics were similar across both groups, although CN patients’ trend to be treated in a lower proportion with angiotensine conversor enzyme inhibitors (ACEi) and mineralocorticoid receptor antagonists (MRA) (43.2% vs 61.4%; p=0.084). According to multimodal congestion assessment, CN patients showed significant higher rates of clinical congestion score (p=0.004), particularly in terms of presence of significant edemas (p=0.023) and jugular vein distension (p=0.001). Doppler vein analysis of suprahepatic veins and portal vein confirmed a significant presence of hepatical congestion (p<0.001 and p<0.001, respectively), and higher VExUS grade (p<0.001) (Supplementary material, Fig. S1. Congestion biomarkers were also higher in CN patients, with median concentrations of NT-proBNP on the admission of 7851 [13063] pg/mL vs 4984 [5935] pg/mL (p<0.025) and median concentrations of CA125 of 47.0 [50] U/mL vs 34.2 [84] U/mL (p-value<0.077). Of note, patients with CN received higher doses of endovenous furosemide during the first 72h after admission (518 [1050] mg vs 280 [453] mg; p=0.010) (Table 2). No significant association was found between CN and long-term outcomes in our cohort.

Among the 100 patients finally included, 28 patients died by day 365, and 18 were re-hospitalized because of a new HF rehospitalization. Patients who had high vascular congestion (VExUS score 2-3) on admission, had a higher proportion of the combined primary end-point (HF rehospitalizations and/or all-cause mortality after 365 days) (p=0.048) and Kaplan-Meier plots in Fig. 2 showed a significant rate too (Long-rank test=0.031). The risk for the primary end-point by day 365 more than doubled in patients with higher VExUS scores on admission (Odds ratio 2.52, 95% CI [1.02–6.26]; p=0.046), although after multivariate adjustment, significance for VExUS score was lost on top of strongest variables (Table 3). In the multivariable analysis, elevated urea levels, larger inferior vena cava diameter and male gender, were identified as independent predictors on top of VExUS score for one-year mortality with an area under the curve for the final model of 0.740 (p<0.001).

Fig. 2.

Survival analysis. Differences between rehospitalization, death composite of events based on VExUS grade and each flow (suprahepatic, portal and renal).

Mortality and readmission rates in HF are generally increasing and risk stratification might be challenging. Torres et al.17 have pointed out the role of ultrasound in assessing prognosis. They identified portal vein pulsatility >50%, monophasic renal pattern and VExUS grade 3 as predictors of in-hospital mortality in patients admitted for HF. Our findings are similar to theirs, showing that VExUS on admission is capable of identifying those patients with higher congestion and worse outcomes in the mid-term. Recently, Espriella et al.12 concluded that discontinuous renal flow identifies higher rates of worsening renal function and increased rates of adverse events. Along the same findings, in the past few months, Sovetova et al.18 published a similar study to ours showing that VExUS grade 3 was strongly associated with worsened renal function, reduced natriuretic response, increased resistance to diuretics and higher incidence of hospital mortality. It also found that monophasic patterns of renal flow and abnormal portal flow had increased risk of adverse events.

However, our data did not support the isolated value of portal flow but renal flow was notably close to reaching statistical significance in the survival regression analysis at one-year follow-up, suggesting a relevant prognostic role that warrants further investigation. Probably this lack of evidence could be due to a limited number of patients with grade 3 alterations, and challenges related to anthropometric factors and the operator-dependent nature of performing ultrasound. However, there is ongoing controversy about the use of renal doppler flow, as some authors argued that this method can be time-consuming,17,19 which contrasts with previous reports highlighting its clinical value. This inconsistency warrants further explanation to better understand its impact and relevance.

Interestingly, no significant association was found between CN and long-term outcomes in our cohort. This divergence from previous studies, such as those by Damman et al.,20,21 who identified CN as a prognostic marker, may stem from differences in patient populations, definitions, or statistical adjustments. Future studies should explore whether tailored interventions targeting CN can modify outcomes.

Multivariate analysis highlights the independent prognostic value of elevated urea levels, increased IVC and male gender in predicting one-year mortality in HF. These findings are consistent with the established role of renal dysfunction and volume overload as key prognostic markers HF. Urea emerged as a significant predictor of outcomes in this study, whereas creatinine did not demonstrate the same prognostic value, which may be attributed to urea’s greater sensitivity to systemic congestion and volume overload. While creatinine is a commonly used marker of renal function, urea levels can reflect changes faster in fluid balance and renal perfusion, as described years ago.22,23 Additionally, male gender was identified as a significant predictor of worse outcomes, potentially due to a higher prevalence of ischemic etiology and comorbidities like diabetes and renal dysfunction.24,25

Interestingly, IVC diameter appears as a significant predictor in the multivariate analysis while VExUS does not, which is controversial given that other studies support VExUS above IVC, as we mentioned before. Maybe it could be explained as the IVC is an objective, unique and easy parameter and VExUS score involves a multiple parameter assessment, which inherently introduces greater variability in the results. Our group is working towards automatic segmentation algorithms to minimize variability in interpretation and reduce these drawbacks.26 Additionally, it should be considered that patients with pulmonary hypertension may present an elevated IVC and this could influence the predictive value of the IVC in our analysis. Previous studies,17 have demonstrated the prognostic value of VExUS; however, our findings support the notion that markers of renal function and congestion continue to play a central role in risk stratification for HF patients. Although VExUS may provide additional prognostic information, its predictive value could be limited by confounding variables and may benefit from being used in conjunction with other clinical parameters to enhance its utility in clinical practice.27

This study’s strengths include its prospective design, the standardized application of the VExUS protocol, and the multiparametric assessment of congestion. However, it has several drawbacks, including the lack of a universal definition for CN, the absence of data on pulmonary hypertension, the degree of right ventricular failure without TAPSE assessment, as well as its single-center design, which limit the generalizability of the results to other healthcare settings.

Future research should explore whether VExUS-guided therapeutic interventions can improve long-term outcomes. Randomized controlled trials investigating VExUS as a decision-making tool for diuretic titration could provide valuable insights.