Acute heart failure (AHF) is a clinical syndrome caused by acute onset or exacerbation of left ventricular dysfunction, leading to decreased myocardial contractility and an increased load on the heart. This syndrome results in an abrupt decrease in cardiac output (CO) and an increase in pulmonary circulation pressure and peripheral vascular resistance, leading to acute pulmonary congestion, pulmonary edema and clinical symptoms that may be accompanied by inadequate tissue and organ perfusion resulting in cardiogenic shock [1, 2]. The etiology of AHF is undoubtedly complex, but is believed to be related to hemodynamic disturbances [3, 4]. A recent report found that the in-hospital mortality rate of AHF was 3%, and the 3 and 5 year mortality rates were 30% and 60%, respectively [5], indicating poor prognosis and high mortality. Pharmacological interventions aimed at achieving rapid decongestion and improving organ perfusion include positive inotropic drugs that increase CO, raising blood pressure, alleviation of tissue hypoperfusion and maintenance of the functions of vital organs [6]. Moreover, the use of diuretics for reducing congestion has been a major treatment mode for AHF patients in clinical practice [7], but about 20% of patients did not exhibit improved symptoms after treatment with diuretic drugs, and furthermore diuretic resistance occurred in $>$30% of them [8, 9].

Ultrafiltration (UF) is an alternative approach to diuretic therapy for congestion management, according to the European Society of Cardiology Guidelines for the diagnosis and treatment of AHF and chronic heart failure [3]. It is advised for patients with obvious volume overload in order to alleviate congestive symptoms and fluid weight (Class IIb, Level of Evidence: B). It has also been found to improve long-term outcomes for patients with acute decompensated heart failure (ADHF) [10]. The use of UF was shown to have little effect on all-cause mortality over the longest follow-up periods studied, but UF reduced all-cause re-hospitalization rates to $\le$30 days and at the longest available follow-up [11], which may be related to the fact that UF therapy could more significantly alleviate volume overload. However, UF may lead to a decline in renal function, mainly manifested as elevated concentrations of serum urea and creatinine, as well as an increased risk of renal failure, bleeding and other complications [12, 13]. Another study also reported that among patients with ADHF who had worsened renal function and persistent congestion, the occurrence of serious adverse events was higher in those receiving UF treatment compared with those treated with pharmacological therapy [14]. Moreover, the duration of UF has been shown to be related to the cost of hospitalization of patients, and if UF is continued after attainment, it can further increase the financial burden on patients. Therefore, it would be of great clinical interest to have a predictor of when a patient has reached the endpoint of UF via hemodynamic changes. Hence, the alternative strategy for monitoring UF attainment by one non-invasive device is an endeavor of practical clinical relevance.

The ultrasonic cardiac output monitor (USCOM) 1A system (USCOM Ltd., Sydney, Australia) is a non-invasive Doppler stroke volume (SV) technique derived from echocardiography, that has been validated for a CO of 0.12 L/min to 18.3 L/min [15]. Moreover, it has the advantages of high repeatability, continuous monitoring and cost-effectiveness, and is especially suitable for predicting hemodynamic changes [16]. Furthermore, there are indicators on USCOM that reflect the volume, such as corrected flow time (FTc) and the systemic vascular resistance index (SVRI) [17, 18]. Compared with traditional invasive monitoring methods, such as pulse indicator continuous CO and Swan-Ganz floating catheters, USCOM can also obtain accurate and reliable data, and has been verified for both adult and pediatric patients [19, 20]. USCOM is easy to operate and trainees reached the same level as trainers after 50 operations, with the learning curve for skill acquisition being significantly shorter [21, 22]. USCOM combined with UF is a valuable tool for cardiologists to diagnose and manage the burden of body fluids, where adjusting UF periods according to USCOM data may well reduce the costs of AHF treatments. However, it remains a challenge to integrate optimally USCOM metrics with UF parameters to determine the exact timing of effective and timely monitoring of the UF endpoint.

The detection of B-type natriuretic peptide (BNP) or N-terminal pro-BNP (NT-proBNP) is recommended for screening, diagnosis and differential diagnosis of AHF, as well as for the assessment of the severity and prognosis of AHF [23, 24]. The variations of BNP before discharge have been independently associated with an increased risk of cardiovascular events, re-hospitalization or death after discharge [25]. At present, the Chinese guideline recommends UF to be applied for 7 consecutive days and BNP concentrations should be reduced by at least 50% compared to baseline [26]. Studies have shown that 30% and 46% declines in BNP at discharge are favorable values for the prognosis of heart failure patients [10, 27]. Other authors proposed a BNP/NT-proBNP reduction $\ge$30% as the standard for effective treatment, and a decrease of BNP/NT-proBNp 30% during hospitalization for AHF as indicative of an increased risk of re-hospitalization and death [28]. Thus, thresholds of a 30% and 50% reduction in BNP concentrations relative to baseline during the UF process were set as criteria to identify one or more potential indicators on the USCOM monitor and to establish a predictive model in the present study.

It would be of great clinical interest to have a predictor of when a patient has reached the endpoint of UF via hemodynamic changes. Therefore, the present single-blind, randomized control trial was designed to evaluate the feasibility of USCOM for determining UF endpoints during AHF treatments and to establish whether it is useful in reducing the financial burden on ADHF patients requiring UF. The primary objective of the trial was to evaluate differences in hemodynamics of patients receiving UF alone (U group) or UF + USCOM (UU group). The secondary objective was to construct a prediction model of potential indicators on USCOM (i.e., inotropy (INO), FTc, SVRI) for achieving UF standards based on threshold criteria of a 30% or 50% decrease in BNP concentration relative to baseline.

The present study found that change rates in MAP, HR, urine output, HCT and BNP over 7 days were similar between the U and UU groups. Our study also evaluated whether USCOM could be used to estimate the endpoints of UF for ADHF patients and provided two UF-endpoint prediction formulae for evaluating a 30% reduction in BNP: –2.462 + 0.028 $\times$ INO – 0.069 $\times$ SVRI, and for a 50% reduction of BNP: –2.640 – 0.088 $\times$ FTc – 0.036 $\times$ HR. Moreover, UF combined with USCOM also reduced the financial burden of treatment and hospitalization for patients.

The introduction of non-invasive devices for monitoring CO in AHF patients represents a significant advance in that they can reduce the occurrence of complications (e.g., infection, thrombosis) compared with invasive hemodynamic monitoring [34]. The present trial also confirmed that USCOM was a safe non-invasive device, as no adverse events due to USCOM were observed within 24 h. Current non-invasive approaches for determining CO include impedance cardiography and echocardiography, but other techniques are under investigation but are proving to have varying efficacies [35]. Operator dependence was undoubtedly the greatest limitation for the application of echocardiography. That is, the operator required advanced experience and echocardiography training, so that the learning curve was even longer compared to that of other non-invasive hemodynamic monitoring approaches [36]. Impedance cardiography was an operator-independent cost-effective and non-invasive approach, but the measurement accuracy might be limited to pathological states, such as too low or high CO values, valvular regurgitation, intracardiac shunts and the incidence of arrhythmia [37]. The clinical application worth of USCOM remains controversial, with its accuracy and precision being assessed with varied results compared with other non-invasive methods [38, 39]. USCOM was easy to operate so that trainees could reach the same level as the trainers after 50 operations, thus the learning curve for skill acquisition was significantly shorter [21]. Moreover, the addition of USCOM also significantly resulted in treatment cost savings and a reduced hospitalization stay length, which confirmed the cost-effectiveness of USCOM as previously reported [16]. Notably, USCOM was also susceptible to operational influences. Consequently, all patients in the UU group in the present trial were monitored by the same skilled operator and three consecutive measurements were made with a deviation of no more than 10% each time, in order to ensure the consistency and reliability of the data.

In the present trial, there were clear differences in the changes of hemodynamic parameters during UF, which reached a steady-state level after 4 days. Since BNP was the standard biomarker, variations of BNP concentrations were associated with the re-hospitalization and mortality rates [26, 28]. Thus, a decrease in the BNP concentration on Day 4 relative to baseline was set as the UF-endpoint to determine its correlation with USCOM parameters. On Day 4, INO and SVRI on USCOM were found to be significantly correlated to a 30% reduction in BNP relative to baseline, and the predictive formula for the UF endpoint of a 30% reduction in BNP was: –2.462 + 0.028 $\times$ INO – 0.069 $\times$ SVRI (AUC ROC 0.831, 95% CI: 0.741–0.920; Hosmer-Lemeshow index 0.814), with a predicted sensitivity, specificity and accuracy of 70%, 83% and 75.0%, respectively, with a cut-off value of 0.567. INO, serving as one of the indicators of cardiac contractility, was also a distinctive hemodynamic parameter of USCOM [40]. Previous studies found that a higher INO could result in a greater SV under the same cardiac preload, while a lower INO was associated with impaired myocardial contractile function [41, 42]. SVRI reflected the cardiac afterload situation, with afterload being another key factor that influenced SV and CO. That is, with a reduction in SVRI, SV and CO were both increased. In the present trial, the elevation trends of INO, SV and CO, and a reduction trend of SVRI, were also observed during the UF process. Previous studies indicated that FTc effectively reflected the cardiac preload situation and was utilized to predict fluid responsiveness [43]. Chaiyakulsil et al. [19] employed USCOM, transthoracic echocardiography and electrical velocimetry for hemodynamic monitoring, and their findings confirmed that the FTc derived from the three non-invasive monitoring approaches could be utilized interchangeably. Moreover, FTc was easily measured by USCOM, and FTc may be a better predictive indicator to assess volume status and diuretic therapy [44]. As FTc was HR dependent, the present trial also confirmed a predictive formula of –2.640 – 0.088 $\times$ FTc – 0.036 $\times$ HR (AUC ROC 0.809, 95% CI: 0.709–0.909; Hosmer-Lemeshow index 0.655) for the UF endpoint of a 50% reduction in BNP, with a predicted cut-off value of 0.572, with sensitivity, specificity and accuracy for this model being 83%, 63% and 72.5%, respectively. Therefore, this trial highlights the feasibility of using USCOM indicators to predict the achievement of UF standards and has certain clinical significance.

However, the kinetics of BNP release vary between individuals and are influenced by factors such as renal function, medication use and comorbidities [45]. In some cases, despite improvements in hemodynamics, there may be a delayed reduction in BNP concentrations. This delay can be attributed to the time needed for the heart to adjust preload and afterload, as well as the clearance rate of BNP from the circulation [46]. In addition, it has also been reported that BNP cannot be used in isolation to measure congestion; rather, the concentration must be assessed in a proper clinical setting, like most other tests, and a precise cut-off point is not suitable. However, as noted by Mueller et al. [47], adjusting UF rates to patients’ vital signs and renal function has been linked to more effective decongestion and fewer heart failure events. It is well known that fluid overload is the main cause of hospitalization for patients with AHF, and changes in urine output may be explored as a predictor of fluid balance [48]. The present trial findings also indicated that urine output was positively correlated with the change rate of FTc (r = 0.255, p = 0.022) and negatively correlated with the VTI change rate (r = –0.27, p = 0.016). Thus, although in this trial, the rate of BNP reduction was used as a criterion for terminating UF, we need to consider other clinical features and the hemodynamic status of patients when considering BNP concentrations, including changes in fluid balance and signs of congestion, to make wise decisions regarding the termination of UF in clinical practice.

Another concern was the influence of comorbidities to hemodynamic changes during UF [49], since diabetes [50], hypertension [51], chronic kidney disease [52, 53] and cardiovascular diseases [54], both demonstrated to have an impact on fluid balance and vascular tone, are crucial factors in hemodynamics [55]. Patients with pre-existing cardiac conditions may be more susceptible to hemodynamic fluctuations during UF [56], and require careful monitoring and management. In contrast, another study demonstrated that the diverse physical conditions of critically ill patients might exert a rather limited influence on the graphic quality of USCOM [57]. In the present trial, the distribution was balanced between the UU group and the U group of patients with ADHF, so that it could be assumed that the hemodynamic impact was the same for both groups during their respective UF sessions.

This trial had several limitations. The selected ADHF patients were from a single center and were not stratified randomly, which might have led to a potential for bias. Besides the relative low number of samples, the frequent comorbidities might have had some influence on the hemodynamic changes in the graphic quality of USCOM and should be validated in a further large cohort trial. Since the kinetics of BNP release vary between individuals and are influenced by various factors, the usefulness of measurements of BNP concentrations may be limited by the possibility that their production and release may lag behind acute changes in hemodynamic measurements, thus additional potential endpoint criteria (e.g., urine output changes) should be explored in the near future. Further research is needed on the applicability of USCOM for different types of AHF patients (e.g., those with chronic kidney disease) and to validate its economic and clinical value after long-term follow-up data have been analyzed.

USCOM data were correlated significantly with UF outcomes and might serve as measures to determine endpoints for congestion therapy with UF in patients with ADHF. UF combined with USCOM also reduced the financial burdens of treatment and hospitalization for patients. No adverse events were reported due to USCOM within 24 h use, indicating that USCOM was a safe and non-invasive monitoring device.

ADHF, acute decompensated heart failure; AHF, acute heart failure; AUC, area under curve; BNP, B-type natriuretic peptide; CI, confidence interval; CO, cardiac output; FTc, corrected flow time; HCT, hematocrit; HR, heart rate; INO, inotropy; IWRS, interactive web response system; MAP, mean arterial pressure; NT-proBNP, N-terminal pro-B-type natriuretic peptide; OR, odds ratio; ROC, receiver operating characteristic; SpO₂, oxygen saturation; SV, stroke volume; SVI, stroke volume index; SVR, systemic vascular resistance; SVRI, systemic vascular resistance index; SVV, stroke volume variation; U, ultrafiltration alone; UF, ultrafiltration; USCOM, ultrasonic cardiac output monitor; UU, ultrafiltration + ultrasonic cardiac output monitor; VTI, velocity time integral.

The datasets used and/or analyzed during the current study are available from the corresponding authors on reasonable request.

YL: Conceptualization, Investigation, Visualization, Writing – original draft. JC: Conceptualization, Formal analysis, Investigation, Visualization, Writing – original draft. JB: Investigation, Visualization, Writing – review and editing. FC: Investigation, Visualization, Writing – review and editing. QW: Conceptualization, Formal analysis, Project administration, Supervision, Visualization, Writing – original draft, Writing – review and editing. FY: Conceptualization, Formal analysis, Project administration, Supervision, Visualization, Writing – original draft, Writing – review and editing. All authors read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

The study adhered to the principles of the Declaration of Helsinki and the protocols were approved by the Institutional Review Board of Tongren Hospital Shanghai Jiao Tong University School of Medicine (approval number: 2021-072-01). Written informed consent was obtained from all enrolled patients.

Not applicable.

This research received no external funding.

The authors declare no conflict of interest.

Supplementary material associated with this article can be found, in the online version, at https://doi.org/10.31083/RCM27100.