Hemodynamic monitoring during veno-venous extracorporeal membrane oxygenation: A scoping review

Abstract

Background

In adult patients receiving veno-venous Extracorporeal Membrane Oxygenation (VV ECMO), cardiovascular performance plays a critical role in determining oxygen delivery, organ perfusion and safe titration of extracorporeal support. Despite the increasing VV ECMO use, contemporary guidance on hemodynamic monitoring remains limited and largely experience-based. This scoping review aimed to map available basic and advanced monitoring approaches and to identify current evidence gaps.

Methods

PubMed, EMBASE, and Cochrane CENTRAL were searched from inception until September 2025, along with reference lists of relevant articles. We included studies of any design reporting techniques, targets, or protocols for hemodynamic monitoring during VV ECMO.

Results

Of 465 records screened, 106 met inclusion criteria. No protocolized, evidence-based hemodynamic monitoring protocol specific to VV ECMO was identified. The available evidence was heterogeneous and mostly derived from physiologic studies or single-center observational cohorts. Findings were narratively synthesized across three domains: basic bedside monitoring, diagnostic/prognostic tools and advanced assessment of cardiopulmonary interaction. Across studies, no monitoring strategy consistently reduced time-to-wean or mortality. Observational data suggested that care bundles and multidisciplinary approaches may reduce complications. However, the risk of bias limits causal inference.

Conclusions

Despite the complex interaction between native cardiovascular function and extracorporeal circulation, VV ECMO lacks consensus on evidence-based hemodynamic monitoring pathways. A pragmatic core monitoring bundle with tiered triggers for escalation is necessary. Future priorities include implementation models based on multidisciplinary teams, specific training, standardized bundles, and multicenter studies aimed to define right ventricular-centered targets to improve safety and clinical decision-making.

Graphical Abstract

Introduction

Over the past decade, the use of veno-venous extracorporeal membrane oxygenation (VV ECMO) for refractory respiratory failure has increased substantially.^1–3 Despite the publication of international guidelines and recommendations,^4,5 considerable variability in VV ECMO management and patients’ outcomes persists across centers. ⁶ While advances have been made in the assessment of lung function and recovery, less attention has been directed toward the cardiopulmonary interactions and hemodynamic abnormalities that frequently occurr during extracorporeal respiratory support. ⁷ VV ECMO directly influences the cardiovascular physiology by reducing the pulmonary vascular resistance, decreasing the right ventricular (RV) afterload, and potentially improving the RV performance. These effects are mediated by correction of severe hypoxemia and hypercapnia, as well as by reductions in intrathoracic pressures resulting from lung-protective ventilation strategies. ⁷ Additional positive consequences may include improved left ventricular (LV) filling pressures. Conversely, the development or progression of cardiac dysfunction, pulmonary deterioration, increased metabolic demand, an imbalance between ECMO flow and native cardiac output (CO), or increased recirculation can impair extracorporeal efficiency and precipitate the cardiocirculatory compromise (Supplemental Figure 1). The COVID-19 experience further highlighted that several etiologies of respiratory failure may severely impact the cardiac function, predisposing patients to different degrees of acute RV and/or LV failure, and in some cases, cardiac arrest.^8–14

Importantly, most of the existing litertaure on hemodynamic monitoring during ECMO combines veno-arterial (VA) and VV configurations, despite their fundamental physiological differences. In VV ECMO, systemic oxygen delivery and extracoporeal efficiency are tightly linked to native CO, pulmonary vascular load, and recirculation, making direct extrapolation from VA ECMO guidance inappropriate.

A VV ECMO-focused synthesis is therefore necessary to support clinically relevant hemodynamic monitoring strategies tailored to this population. Such strategies should encompass basic “mandatory” monitoring tools,a s well as more “advanced” techniques, adjusted according to patient-specific factors, disease evolution and the occurrence of complications. Given the limited volume and heterogeneous nature of the available literature, a scoping review methodology was selected to comprehensively map current practices and identify knowledge gaps.

Methods

This scoping review aimed to map the available literature and summarize key concepts related to hemodynamic monitoring during VV ECMO as identified by the COVID-19 Critical Care Consortium (COVID-Critical). The specific objectives were:

1. To describe existing evidence on hemodynamic monitoring techniques in adult VV ECMO patients

2. To propose a practical, multi-step approach for clinical practice

An electronic search of PubMed, EMBASE, and Cochrane CENTRAL was conducted from database inception through September 2025, using a combination of controlled vocabulary and free-text terms related to VV ECMO and hemodynamic or cardiovascular monitoring. Key search terms included: “veno-venous extracorporeal membrane oxygenation” and “VV ECMO” combined with “ hemodynamic monitoring,” “cardiovascular monitoring,” “cardiac output,” “right ventricular,” “echocardiography,” “pulmonary artery catheter,” “ microcirculation” and “recirculation.” Reference lists of included articles and relevant reviews were hand-searched to identify additional studies.

We included studies involving adult patients (≥18 years) supported with VV ECMO, in which the primary focus was hemodynamic monitoring, cardio-pulmonary interactions or cardiovascular assessment during VV ECMO. Across included sources there is a broad agreement that hemodynamic monitoring is clinically essential during VV ECMO; however, there is no consensus on a standardized VV ECMO specific monitoring protocol, with limited structured recommandations and strategies vary between centers.

All study designs were eligible, reflecting the exploratory nature of this scoping review. Studies focused on VA ECMO or hybrid configurations whitout VV-specific data, as well as pediatric-only studies, were excluded while studies not directly involving patients (i.e., ex-vivo device testing) were included only if they addressed questions relevant to the defined patients population.

The search was not limited but mainly focused on English language studies. Two reviewers (MEDP, SM) independently extracted data using a standardized template capturing author/year, setting, design, population/sample size, ECMO configuration, monitoring modality and key findings. Disagreements were resolved by consensus with a third investigator (RL). Data were stored in a structured table, synthesized descriptively and mapped to predefined monitoring domains (Table 1). Analysis was descriptive and was performed according to the PRISMA extension for scoping review (Figure 1; Supplemental Table 1).

Table 1.

Cardiocirculatory monitoring during VV ECMO.

Monitoring domain	Modality	Invasive	ECMO-specific interpretability	Assessment frequency a	Key ECMO-specific considerations
Basic monitoring	ECG	No	Preserved	Continuous	Arrhythmias may reflect hypoxia, RV strain, electrolyte imbalance, or cannula irritation
	Pulse oximetry	No	Context-dependent	Continuous	Influenced by ECMO flow/native CO ratio, recirculation, peripheral vasoconstriction
	Urine output	Yes	Preserved	Hourly	Reflects global perfusion but insensitive to early shock
	Capnography	No	Limited	Continuous	End-tidal CO2 decoupled from PaCO2 during ECMO
	Capillary refill time (CRT)	No	Preserved	Daily + after instability	Surrogate of peripheral perfusion; interpretation similar to non-ECMO patients
	Mottling score	No	Preserved	Daily + after instability	Reflects microcirculatory dysfunction; may persist despite normalized macro-hemodynamics
	Invasive arterial pressure	Yes	Preserved	Continuous	Waveform damping common with low pulsatility
	Arterial blood gases	Yes	Preserved	After ventilatory or ECMO changes	PaO2 reflects ECMO/native flow balance, not lung function alone
	Pre-oxygenator venous saturation	Yes	Conditional	Serial trending + after changes	Strongly influenced by cannula position, recirculation, ECMO flow
	Chest X-ray	No	Indirect	Daily + after changes	Cannula position, pulmonary complications
	POCUS (focused cardiac/lung)	No	Preserved	Daily + after instability	Used for RV size/function, volume status, cannula position
	Near-infrared spectroscopy (NIRS)	No	Indirect	Continuous	Reflects regional perfusion trends; not a direct CO surrogate
	Lactate	Yes	Indirect	After instability	Late and nonspecific marker; trends more informative than single values
	ECMO flow & pressures	No	Preserved	Continuous	Drainage pressure sensitive to preload and cannula position
	Circuit inspection & recirculation	No	Preserved	Daily	Visual and physiological indicators of ECMO efficiency
Diagnostic/prognostic monitoring	Bilirubin & liver enzymes	Yes	Indirect	Daily	Reflect hepatic congestion, hypoperfusion, hemolysis
	BNP/NT-proBNP	Yes	Context-dependent	Baseline + trends	Reflect RV strain, volume overload, disease severity
	Troponin	Yes	Context-dependent	Baseline + trends	Marker of myocardial injury, prognostic rather than diagnostic
	IL-6	Yes	Prognostic	Selected cases	Inflammatory severity; not hemodynamic per se
Advanced monitoring	Transthoracic echocardiography (TTE)	No	Preserved	Serial	RV-pulmonary coupling, septal position, cannula effects
	Transesophageal echocardiography (TEE)	Yes	Preserved	Targeted	Preferred for cannula position, RV dysfunction, septal shift
	Pulmonary artery catheter	Yes	Conditional	Selected patients	CO and filling pressures affected by ECMO flow and recirculation
	Pulse contour analysis	No	Limited	Continuous	Unreliable with low pulsatility, RV failure, low tidal volume
	Transpulmonary thermodilution	Yes	Limited	Selected patients	Indicator loss into ECMO circuit limits accuracy
	Sublingual microcirculation	No	Emerging	Selected expert centers	Assesses micro-macro coherence; not invasive

^aSuggested frequencies ECG: Electrocardiogram; POCUS: Point of care ultrasound; BNP: B-type natriuretic peptide; NT-proBNP: N-terminal pro-B-type natriuretic peptide; RV: right ventricle; ECMO: extracorporeal membrane oxygenation; CO: cardiac output; CO₂: carbon dioxide; PaCO₂: partial pressure of carbon dioxide in arterial blood.

Figure 1.

PRISMA_ScR flow diagram.

Monitoring modalities were categorized into three tiers: basic (low-risk procedures, bedside screening measures, applicable in all patients), diagnostic/prognostic (tools reflecting the evolving severity of organ dysfunction and supporting escalation), and advanced (specialized or invasive modalities providing detailed quantification of cardiopulmonary interaction, CO, or microcirculation). Categorization was based on feasibility, invasiveness, interpretative complexity and common clinical use in VV ECMO. Several modalities, including point of care ultrasound (POCUS) and CO estimation, were recognized to span tiers depending on clinical context and depth of assessment.

Results

The electronic search retrieved 465 citations. After removal of duplicates and application of eligibility criteria, 106 articles were included in the final analysis. (Supplemental Table 2). Most of the identified articles addressed a single hemodynamic monitoring modality. The key findings were categorized into three domains: (1) basic, (2) diagnostic/prognostic, and (3) advanced monitoring tools (Table 1). Individual sources of evidence were mapped by monitoring domain and linked to the review objectives. Supplemental Table 2 provides study-level characteristics and key findings, while Supplemental Table 3 links monitoring modalities to Objective 1 (evidence on monitoring techniques) and Objectives 2 (development of the proposed multi-step practical approach)

Basic hemodynamic monitoring

Basic monitoring includes clinical examination, blood pressure, heart rate, pulse-oximetry, chest radiography, and urine output. ¹⁵

Capillary refill time (CRT)

CRT is a rapid metric reflecting peripheral perfusion and is easily performed and interpreted. Pressure is applied to the ventral surface of a distal phalanx until blanching occurs, maintained for 10 s, and then released. The time in seconds that elapses before reperfusion reflects circulatory status. A CRT of less than 3 s is generally considered normal. The ANDROMEDA-SHOCK trial demonstrated that CRT utility extends beyond diagnostic purposes and may be used to guide fluid resuscitation strategies in unstable patients. ¹⁶

Mottling score

Microcirculation blood flow impairment may be pronounced in critically ill patients and may persist despite normalization of macro-hemodynamic variables (macro-microcirculatory dissociation). ¹⁷ Mottling is a patchy skin discoloration reflecting cutaneous hypoperfusion, often related to local vasoconstriction and endothelial dysfunction. The mottling score provides a semi-quantitative assessment based on the extent of patchy skin discoloration around the knee and lower limb. Increasing extent reflects a worsening of the cutaneous hypoperfusion and microcirculatory impairment.^18–21

Arterial blood waveform

The arterial waveform provides a real-time signal of the arterial blood pressure and enables derived indices such as the pulse pressure variation (PPV). When appropriately measured and interpreted, PPV can help predict fluid responsiveness. ²² However, reliability is limited by waveform quality and is reduced in conditions commonly encountered in Acute Respiratory Distress Syndrome (ARDS) and VV ECMO, such as peripheral vasoconstriction, low tidal volume (Vt), reduced low lung compliance, dysrhythmias, and increased intra-abdominal pressure. No randomized controlled trials have compared PPV-guided fluid management with standard care in VV ECMO. However, a dynamic assessment such as a transient tidal volume challenge for approximately 2 min may improve interpretability in selected cases. ²³ Passive Leg Raising (PLR) is an alternative dynamic test based on a preload redistribution that is less dependent on tidal volume, respiratory rate or pulmonary compliance. PLR mimics fluid expansion by shifting blood from lower limbs and splanchnic compartment. ¹⁹ A CO increase >10% after PLR suggests fluid responsiveness and may identify patients likely to benefit from fluid loading. ¹⁹

Arterial and venous blood gas analysis

An arterial line is essential to measure arterial oxygen content (CaO₂), partial pressure of arterial oxygen (PaO₂), and carbon dioxide (PaCO₂). CaO₂ depends on hemoglobin concentration and ECMO blood flow, which determines oxygen exchange within the membrane lung. PaCO₂ is primarly regulated by sweep gas flow and blood flow, in combination with patient-related factors such as minute ventilation. In VV ECMO, arterial oxygenation reflects mixing between oxygenated extracorporeal blood and native venous blood and is influenced by ECMO blood flow, recirculation fraction, and native CO. ²⁴ The site of arterial blood to monitor oxygenation in VV ECMO is required to be as far as possible from the reinjection site like in VA ECMO.

Venous blood gas analysis complements arterial sampling and reflects the mixed venous blood from all body regions and CO₂ clearance. ⁷ In VV ECMO, interpretation depend on cannula configuration (jugular-femoral, femoro-jugular, femoro-femoral, or double-lumen) that significantly influences the anatomical site of venous blood mixing. For this reason, venous sample should be obtained upstream of the membrane lung. ²⁵ Reliable interpretation requires stable ECMO flow and oxygenation without fluctuations in drainage, stable metabolic demand (e.g. absence of shivering), stable native CO and RV function, known recirculation fraction, appropriate cannula position, stable intra-thoracic, intra-cardiac, and intra-abdominal pressures, and no recent major variation in ventilator settings, blood transfusions, and vasoactive support. ²²

DO₂, VO_2, and PCO₂ gap

Simultaneous arterial and venous sampling allows calculation of oxygen delivery (DO₂), oxygen consumption (VO₂), and DO₂/VO₂ ratio. These parameters provide only indirect information on oxygen supply adequacy, require accurate CO meausurements and should be interpreted alongside recirculation and clinical context.

The PCO₂ gap, defined as the difference between venous and arterial PCO₂, is proportional to carbon dioxide production (VCO₂) and reflects the adequacy of the microcirculatory blood flow. A PCO₂ gap >6 mmHg suggests inadequate CO. Persistent elevation has been associated with worse outcomes and reflects impaired venous return from poorly perfused capillary beds. ²⁶ Changes in the PCO₂ gap, occur earlier than lactate variations and may represent a signal of CO impairment. ²⁷

Serum lactate measurement

Lactate is a metabolic product of anaerobic glycolysis and may indicate inadequate DO₂. Nevertheless, it is not specific and may also rise due to catecholamine administration or reduced hepatic clearance. It is generally a later marker of hypoperfusion and may be less sensitive than SvO2 or the PCO₂ gap. ²⁸ Serial measurements and lactate clearance provide a surrogate for the magnitude and duration of global tissue hypoxia and are more reliable for risk stratification than isolated values, particularly within prognostic ECMO scores. ²⁹

Near infrared spectroscopy (NIRS)

Near infrared spectroscopy (NIRS) provides continuous monitoring of regional oxygen saturation (rSO₂), reflecting the interaction between macro-circulation and organ-specific auto-regulation, and serving as an indirect indicator of peripheral DO₂.^30,31 NIRS cannot distinguish between reduced oxygen delivery and increased oxygen consumption and could be consider as an adjunct regional perfusion monitor. Hypoperfusion is commonly defined as rSO₂ <50% or a decrease >20% from the baseline. Serial measurements, rather than absolute values, ³⁰ are more informative for detecting hemodynamic deterioration. In ECMO patients, NIRS is primarily used for cerebral monitoring to detect hypoxic-ischemic injury.^32–34 It may also identify cerebral vasoconstriction related to rapid PaCO₂ changes, particularly during the first 24 h of ECMO, potentially facilitating controlled PaCO₂ correction and reducing the need for frequent blood gas analyses.

Point-of-care ultrasound (POCUS)

POCUS is widely used in critical care medicine and enables goal-directed, and dynamic assessments. Thoracic ultrasound can identify bilateral B-lines, non-uniform lung aeration, pleural lines abnormalities, reduced lung sliding, and consolidation patterns. In VV ECMO, POCUS assists in evaluating vascular patency, cannula position, RV/LV size and function, intravascular volume status, lung aeration, and potential causes of shock. ⁶ This technique provides important screening information and can guide escalation to comprehensive echocardiography when abnormalities are detected.

ECMO circuit monitoring

Variations in ECMO blood flow can reflect hypovolemia or vasodilation. Excessive negative drainage pressures (“chattering”) may indicate inferior vena cava collapse due to excessive suction and/or insufficient venous capacitance. Causes include hypovolemia, vasodilatation, coughing, Valsalva maneuvers, inflow obstruction, or an undersized drainage cannula. Excess negative drainage pressure may result from high pump speed relative to venous inflow resistanceRoutine inspection of cannulas, pump, tubing, and mebrane lung is therefore considered an integral component of basic VV ECMO monitoring to ensure integrity and patient’s safety.

Monitoring recirculation fraction

Recirculation is the fraction of oxygenated blood reinfused into the right atrium that is immediately withdrawn back into the drainage cannula.^35,36 In VV ECMO settings, an increasing recirculation reduces the effective gas exchange. Native CO significantly influences recirculation as higher fractions may occur in RV failure and low CO states. Recirculation is affected also by pump speed, ECMO blood flow, intrathoracic, intracardiac and intra-abdominal pressures, cannula type, size, and position.

Quantification is challenging and thermodilution or ultrasonic transit-time methods have been described. Recirculation can be estimated using blood gas-derived formulas:

$\mathit{R}=\left({\mathit{SO}}_2\hspace{0.17em}\mathit{preoxy}-{\mathit{SvO}}_2\right)\hspace{0.17em}/\hspace{0.17em}\left({\text{SO}}_2\hspace{0.17em}\mathit{postoxy}-{\mathit{SvO}}_2\right)$

SO₂ preoxy: is the saturation of blood in the venous cannula entering the oxygenator.

SO₂ postoxy: is the saturation of blood in the arterial cannula, leaving the oxygenator.

SvO₂: is the mixed venous saturation.

However, this method is limited because accurate SvO₂ measurement often requires sweep gas interruption and ventilator adjustments, making routine application impractical.

Diagnostic and prognostic perfusion and cardiovascular monitoring

Bilirubin and liver enzymes

Hepatic injury during VV ECMO may reflect hemodynamic alterations, venous congestion, and reduced DO₂. ³⁷ Liver dysfunction can be detected using markers of hepatocellular injury (aminotransferases), hepatic clearance and biliary secretion (bilirubin), and synthetic function (prothrombin time and albumin). Hypoxic hepatitis is characterized by marked increases in aminotransferases and occurs in the setting of respiratory failure, shock, or cardiac failure ³⁸ and may involve hepatic ischemia and venous congestion from elevated central venous pressure and RV failure. ³⁹ In addition, many medications commonly used in ARDS management (e.g.antibiotics, corticosteroids, catecholamines) are potentially hepatotoxic. Elevated liver enzymes are frequently observed in ECMO patients and are associated with disease severity, systemic inflammation and hemolysis. Hyperbilirubinemia is common during ECMO, with a reported incidence of 20–40%, and is primarly related to systemic hypoperfusion and cardiac failure. ⁴⁰ As a marker of hepatic dysfunction, bilirubin is incorporated into prognostic scoring systems for critically ill patients. ⁴¹ Therefore, regular monitoring is recommended, and serial trends may help identify evolving cardio-circulatory compromise when alternative causes of liver injury are excluded.

BNP, NT-pro-BNP and troponin

Biomarkers of cardiac myocyte stretch, such as brain natriuretic peptide (BNP) and N-terminal-natriuretic-peptide (NT-proBNP), are well established for the diagnosis and prognosis of heart failure. ⁴² In critically ill patients, elevations may reflect the global severity of the circulatory compromise rather than an isolated myocardial disease. Higher levels (>800 pg/mL) are associated with infection, organ dysfunction, and overall disease severity, representing an independent predictor of mortality. Cardiac troponins (T and I) can also have prognostic value in pulmonary diseases, including pneumonia and chronic obstructive pulmonary disease. ⁴³ Serial measurement of cardiac biomarkers during ECMO may help identify evolving cardio-circulatory injury or increased mortality risk, although their prognostic value appears lower than lactates in some cohorts.^41,44

Advanced cardio-circulatory monitoring

Transthoracic and transesophageal echocardiography

Echocardiography represents the cornerstone of advanced hemodynanic monitoring during VV ECMO, as it integrates assessment of cardiac structures, ventricular function, pulmonary vascular load, and the dynamic interaction between the native circulation and extracorporeal support. This modality could be considered the primary tool to identify RV dysfunction, guide ventilatory and ECMO adjustments and evaluate causes of persistent hypoxemia or hemodynamic instability despite adequate extracorporeal blood flow.

The RV has emerged as the central determinant of hemodynamic stability and ECMO efficiency in patient supported with VV ECMO. Several observational studies highlighted that acute RV dilatation, reduced systolic function and interventricular septal flattening are common during severe ARDS and are independently associated with adverse outcomes.^45–48 RV end-diastolic area, RV/LV area ratio, RV fractional area changes, tricuspid anular plane systolic excursion (TAPSE) and tricuspid anular S′ velocity provide complementary information on RV size and contractile performance. These parameters are particularly relevant during VV ECMO since RV failure may reduce pulmonary blood flow, impair systemic oxygenation delivery and increase recirculation fraction despite high extracorporeal flows. Assessment of pulmonary vascular load during VV ECMO could be determined through evaluation of tricuspid regurgitation velocity, pulmonary artery doppler acceleration time and paradoxical septal motion. All these parameters were frequently described as markers of elevated pulmonary vascular resistence and RV-pulmonary artery uncoupling. These findings support clinical decisions regarding ventilatory settings, fluid management and the need of adjunctive therapies aimed at RV unloading.

Although VV ECMO primarly supports gas exchange, LV performance remains critical for systemic oxygen delivery. LV underfilling is common during VV ECMO as a consequence of reduced preload, high intrathoracic pressures and RV dysfunction. Measurements of LV outflow tract velocity-time integral (LVOT-VTI) are repeatedly used as a surrogate of native CO and to explain inadequate oxygen delivery or persistent hyperlactatemia despite adequate ECMO gas exchange. Reduced LV ejection fraction or new regional wall-motion abnormalities were described in the context of sepsis-related cardiomyopathy, myocardial ischemia, or COVID-19 associated myocardial injury, and could significantly alter the balance between oxygen delivery and consumption.^49,50

Echocardiography also plays a key role in evaluating interpretation-dependent variables such as preload and venous congestion. Inferior and superior vena cava size and respiratory variation are commonly used to support volume management decisions. Right atrial dilatation and dilated coronary sinus are described as indirect markers of elevated right-side filling pressures and venous congestion, which can compromise ECMO drainage and organ perfusion (Table 2, Supplemental Table 4).

Table 2.

Echocardiographic assessment during veno-venous ECMO (VV ECMO).

Cardiac structure	Measurement	VV-ECMO–specific relevance and interpretation
Left ventricle (LV)	Morphology (size, wall thickness)	LV underfilling is common during VV ECMO due to reduced preload and high intrathoracic pressures; small LV size may indicate hypovolemia or RV failure–related preload limitation
	Ejection fraction/FAC	Reduced LV systolic function may reflect sepsis-related cardiomyopathy, myocardial injury (e.g., COVID-19), or ventricular interdependence due to RV pressure overload
	Wall motion abnormalities	Suggest ischemia, myocarditis, or stress cardiomyopathy; may impair systemic oxygen delivery despite adequate ECMO gas exchange
	Mitral annular S′ wave	Useful when EF is limited by loading conditions; reduced values suggest impaired longitudinal LV function
	LVOT velocity–time integral (VTI)	Surrogate of native cardiac output; low VTI may explain inadequate systemic oxygen delivery or high recirculation despite adequate ECMO flow
LV diastolic function	E/A ratio; E/e′ at mitral annulus	Elevated filling pressures may worsen pulmonary congestion and RV afterload; interpretation must account for ventilation and ECMO-related preload changes
	Left atrial volume index	Chronic marker of elevated LV filling pressure; helps differentiate acute preload limitation from chronic diastolic dysfunction
Left atrium	Size and volume	Enlargement suggests pre-existing diastolic dysfunction that may limit tolerance of fluid loading during VV ECMO
Valvular assessment (left heart)	Aortic/mitral regurgitation or stenosis	Significant regurgitation alters effective forward flow and oxygen delivery; impacts interpretation of CO and ECMO flow adequacy
Right ventricle (RV)	RV end-diastolic area/LV end-diastolic area ratio	Key marker of RV dilation; RV/LV ratio >0.6–1.0 suggests RV pressure/volume overload and is associated with worse outcomes during VV ECMO
	RV shape (triangular vs rounded apex)	Rounded apex indicates RV pressure overload and acute cor pulmonale
	RV wall thickness	Increased thickness suggests chronic pulmonary hypertension; influences reversibility of RV dysfunction
	McConnell’s sign	May indicate acute RV dysfunction due to increased pulmonary vascular resistance
RV systolic function	TAPSE, tricuspid annular S′ wave	Reduced values indicate RV systolic dysfunction; central determinant of pulmonary blood flow, ECMO efficiency, and recirculation
	RV fractional area change (FAC)	Quantitative assessment of RV systolic performance; low FAC associated with adverse outcomes
	Pulmonary artery Doppler (acceleration time)	Shortened acceleration time suggests elevated pulmonary artery pressures and increased RV afterload
RV diastolic function	E/A trans-tricuspid flow	Impaired RV filling may worsen venous congestion and limit ECMO drainage
Interventricular septum	Paradoxical septal motion	Indicates RV pressure overload and ventricular interdependence, potentially reducing LV preload and systemic output
	Eccentricity index	Quantifies septal flattening; correlates with RV afterload and severity of pulmonary hypertension
Tricuspid regurgitation	TR velocity/estimated RV systolic pressure	Elevated RVSP reflects increased pulmonary vascular resistance; important target during VV ECMO management
Valvular assessment (right heart)	Tricuspid regurgitation severity	Severe TR exacerbates venous congestion and reduces effective pulmonary flow
Right atrium	Size and volume	RA dilation reflects chronic or acute RV failure; contributes to impaired venous return and ECMO drainage instability
Interatrial septum	Patent foramen ovale (color Doppler, bubble study)	Right-to-left shunt may worsen hypoxemia and reduce VV ECMO efficiency
Venous system (TEE)	IVC/SVC size and respiratory variation	Guides preload assessment and fluid responsiveness; limited by ECMO suction and ventilation
	Dilated coronary sinus	Suggests elevated right-sided filling pressures
Vascular assessment	Cannula-related thrombosis/stenosis; aortic dissection; severe atheroma	Identifies ECMO-related complications affecting flow, oxygen delivery, and embolic risk

E/A: early diastolic peak velocity/atrial contraction velocity; E/e′: early diastolic transmitral velocity/early diastolic tissue Doppler velocity; FAC: fractional area change; IVC: inferior vena cava; LV: left ventricle; LVOT: left ventricular outflow tract; RV: right ventricle; SVC: superior vena cava; TAPSE: tricuspid anular plane systolic excursion; TEE: transesophageal echocardiography.

Transthoracic echocardiography is generally preferred for serial bedside assessments, while transesophageal echocardiography is particularly valuable when acoustic windows are limited, during proning position or when detailed evaluation of cannula position, atrial spetum or pulmonary artery flow is required. This scoping review did not identify evidence favoring one modality over the other. Rather, the choiche should be individualized based on clinical context, patient’s characteristics and operator’s expertise.

Pulmonary artery catheter (PAC)

In selected cases, PAC could be used to provide measurements of right-sided pressures, pulmonary artery pressures (PAP), pulmonary artery occlusive pressure (PAOP), and mixed venous saturation. Systemic and pulmonary vascular resistances are calculated from these variables, and CO is measured using a thermodilution method. PAC-derived data may be useful to assess adequacy of ventricular support, diagnose ECMO-related complications (e.g., thrombosis) and guide management in refractory shock or suspected RV failure. However, thermodilution-based CO measurements might be affected by ECMO flow, variations in venous suction, cannula position, RV contractility, recirculation, systemic and pulmonary vascular tones, and any coexisting valvular disease. Pulmonary artery pressures are generally reliable, except in the presence of intracardiac shunt defects and must be interpreted in the context of intrathoracic pressures. Consequently, trends analysis is often more usefult than absolute values. PAC placement carries risks, including tachy-dysrhythmia, pulmonary artery rupture, catheter knotting, vascular injury (e.g.carotid artery damage or arterio-venous fistula), valvular damage, and infection. ⁵¹

Although PAC use remains common in mechanical circulatory support, the routine application in VV settings has declined over the past decade as no consensus exists regarding its systematic use. ⁵² Nevertheless, recent evidence suggestes that PAC may retain a key role in patients with refractory shock associated with RV dysfunction and/or severe ARDS, ⁵³ where abnormal expansion of West zones 1 and 2 may occur due to elevated transpulmonary pressure, particularly in hypovolemic status. In such cases, the transpulmonary pressure gradient (mean PAP-PAOP) remains useful for assessing pulmonary vascular abnormalities. ⁵⁴

Pulse contour wave analysis

Pulse contour wave analysis estimates CO by analyzing the systolic portion of the arterial pressure waveform, which can be influenced by stroke volume, vascular compliance, aortic impedance, and peripheral vascular resistance. This technique provides continuous CO monitoring and dynamic analysis of fluid responsiveness, including stroke volume variation and systolic pressure variation between the inspiratory and expiratory phase of mechanical ventilation. ⁵⁵ Several devices combine pulse contour analysis with thermodilution calibration but accuracy may be affected by arrhythmias, RV failure, spontaneous breathing, low Vt, and drainage cannula position. Indicator dilution calibration may be inaccurate results for redistribution of the indicator into the ECMO circuit, although accuracy improves, when ECMO flows are minimal.

Sublingual microcirculation

The microcirculation represents the terminal vascular network responsible for oxygen delivery to tissues and is critical for organ function. Assessment of sublingual microcirculation has emerged as a potential tool for guiding clinical decisions during ECMO.^55,56 Key bedside indices include total vessel density, perfused vessel density, microvascular flow index, percentage of perfused vessels, and flow heterogeneity index. Advances in handheld vital microscopy and automated analysis software enable bedside visualization and quantitative assessment of microcirculatory flow. ⁵⁷ Direct microcirculatory visualization may identify patients at risk despite normalization of macro-hemodynamic variables and support the development of microcirculatory-guided resuscitation strategies in conjunction with systemic hemodynamic goals.

In VV ECMO, available evidence remains limited, but sublingual microcirculatory alterations have been described in COVID VV ECMO cohorts and may persist despite macro-hemodynamic normalization.^56–58

Multi-step practical approach

Hemodynamic responses during VV ECMO are complex and highly variable, influenced by microthrombosis, pulmonary vascular remodeling, hypoxic vasoconstriction, acidosis, inflammation, mechanical ventilation-induced RV strain, and sepsis-related metabolic demands.^59,60 Moreover, RV dysfunction may be exacerbated by ECMO-related preload and afterload changes. ⁶¹ Hemodynamic monitoring is, therefore, essential to optimize perfusion, improve gas exchange, and minimize ventilator-induced lung injury.

Three interrelated monitoring dimensions are proposed:

1. Basic mandatory monitoring: it provides information on systemic perfusion and intravascular volume status through clinical examination, arterial waveform analysis, urine output monitoring, tissue oxygenation markers (lactate, blood gases, rSO₂,), POCUS, ECMO circuit assessment, and screening echocardiography.

2. Diagnostic/Prognostic monitoring: it identifies disease progression and evolving cardiac dysfunction evaluating early warning signs in terms of organ dysfunction. It supports timely escalation when clinical deterioration or discordant signals emerge.

3. Advanced monitoring: it provides detalied assessment of native cardiac function, cardiopulmonary interactions, macro- and micro-circulation coherence using comprehensive echocardiography, PAC, transpulmonary thermodilution/pulse contour methods, and microcirculatory imaging.

This practical approach, categorized by theme, has been summarized in Table 1 and Figure 2.

Figure 2.

Summary of general principles for recommended hemodynamic monitoring in patients undergoing veno-venous extracorporeal membrane oxygenation. AST: aspartate aminotransferase; ALT: alanine aminotransferase; COVID-19: coronavirus disease 2019; ECMO: extracorporeal membrane oxygenation; NIRS: near infrared spectroscopy; POCUS: point of care ultrasound; TEE: transesophageal echocardiography; TTE: transthoracic echocardiogram.

It is imperative that clinicians have a thorough understanding of limitations, accuracy, and risks of the each method to avoid generating spurious data. Perhaps, the greater challenge lies in “understanding the number” once it is generated. Therefore, we suggest that interpretation of hemodynamic parameters should integrate four aspects: quantitative and qualitative elements, temporal trends, global metabolic context including the balance between oxygen delivery and consumption, and the contributions of both ECMO support and patient physiology.

Challenges in cardio-circulatory monitoring

Besides specific etiologies requiring targeted management, certain clinical scenarios warrant special attention during VV ECMO. Prone position is commonly used in ARDS and RV failure and, in some cases, may reduce ventral-dorsal trans-pulmonary pressure gradients, promoting more homogeneous perfusion and recruitment of collapsed lung alveoli.^62–64 It might also be beneficial during ECMO ⁶⁵ by lowering airway pressures, reducing hypercapnia, unloading the RV, improving venous return, and increasing CO.^66–69 Despite these advantages, it may be less effective in advanced disease stages and may increase the risk of arrhythmias, cannulas displacement and cardiac arrest. Serial cardiac assessment is therefore essential, particularly to detect reductions in cardiac preload. Bedside echocardiography allows rapid, dynamic evaluation of RV function using dedicated views (apical four-chamber, apical long-axis view), ⁷⁰ and assessment of CO response to therapeutic interventions. Unlike other monitoring methods, echocardiography is not significantly affected by changes in body pressure distribution or thoraco-pelvic supports that could increase intrathoracic pressures. Prone positioning is tipically performed with the patient on the stomach with the left arm raised above the head and the right arm aligned with the torso. A pillow placed beneath the left arm slightly elevates the left hemithorax, facilitating echography transducer positioning (Supplemental Figure 2). Due to its feasibility and repeatability, echocardiography could be performed during each prone session to directly evaluate the effect of variations in ventilatory settings (i.e. recruitment maneuvers) and should be incorporated into a multimodality monitoring strategy during prone position.

Another scenario requiring meticulous cardiovascular monitoring is the awake-ECMO strategy.^71–73 Increased work of breathing, discomfort, pain, and anxiety may lead to high oxygen consumption and CO₂ production. Hemodynamic monitoring remains similar to sedated patients but maintaining control of volume status, preload, and venous return may be more challenging. Prolonged negative intrathoracic pressures generated by spontaneous respiratory efforts, associated with metabolic interactions, can affect extracorporeal gas exchange efficiency.^74,75

Discussion

This scoping review provides a VV ECMO-focused synthesis of hemodynamic monitoring that addresses important limitations of the existing literature. Many prior reviews and recommendations on ECMO hemodynamic assessment combine VA and VV configurations despite their fundamentally different physiological objectives and monitoring priorities. In VV ECMO, extracorporeal support is enterily dependent on the native CO and pulmonary circulation, and systemic oxygen delivery is governed by the interaction between RV function, pulmonary vascular load, recirculation and ventilatory mechanics. Consequently, monitoring strategies derived from VA ECMO, where circulatory support is provided directly, are not readly transferable to VV practice. By restricting inclusion to adult VV populations and explicitly mapping monitoring modalities according to feasibility, invasiveness and clinical intent, this review clarifies which tools have been described, how they are applied and where evidence remains sparse or inconsistent. A scoping review methodology was chosen because the available evidence is heterogeneous, predominantly physiologic or observational and rarely structured around protocolized monitoring pathways.^76,77 The literature encompasses diverse study designs with limited comparability across populations and monitoring techniques. Instead a scoping approach allows comprehensive mapping of evidence landscape, identification of conceptual patterns (e.g. RV-centered monitoring, reciruclation assessment) and explicit delineation of knowledge gaps. The mapped literature demonstrates that monitoring strategies are primarly used to support early detection of complications, guide ECMO and ventilator titration, support in troubleshooting, monitor circuit-patient interactions and support a multidisciplinary decision-making based on cardio-pulmonary physiology. This process is highly relevant to VV ECMO management but is not consistently captured in traditional mortality-focused analyses. However, these outcomes are particularly relevant to perfusion practice and to prevention of avoidable complications. By synthesizing this evidence and proposing a pragmatic, tiered monitoring framework with three step approach, this scoping review provides a structured foundation for future hypothesis-driven studies and protocol development specific to VV ECMO, shifting the focus from hypoxemia severity alone toward recognition of RV dysfunction and pulmonary vascular abnormalities.^78–80

Gaps of knowledge and future directions

This scoping review identified substantial variability in monitoring strategies and significant knowledge gaps, with few studies outlining clear management protocols for VV ECMO patients. However, several future directions can be defined including standardized evaluation of monitoring modalitie, multicenter protocols to address current inconsistencies, development of less invasive and more reliable dynamic monitoring techniques, and application of artificial intelligence and machine learning algorithms to high-quality datasets for early prediction of hemodynamic deterioration. Automated feedback loops may eventually allow real-time titration of ECMO support based on patient-specific physiological demands. Finally, future studies should assess whether protocolized monitoring strategies improve clinically relevant outcomes beyond mortality, including complication detection and decision support.

Conclusions

Understanding the complex interplay between ventilation, hemodynamics, and metabolic parameters is fundamental in patients with respiratory failure supported by VV ECMO. At present, no formal guidelines define optimal hemodynamic monitoring strategies in VV ECMO. Accordingly, we propose a sequential, patient-centered approach integrating basic diagnostic/prognostic and advanced tools. Despite progress, key questions remain unanswered, underscoring the need for dedicated guidelines and continued technological innovation to support clinicians and improve patient outcomes.

Supplemental material

Supplemental Material - Hemodynamic monitoring during veno-venous extracorporeal membrane oxygenation: A scoping review

Supplemental Material for Hemodynamic monitoring during veno-venous extracorporeal membrane oxygenation: A scoping review by Roberto Lorusso, Maria Elena De Piero, Silvia Mariani, Justine M. Ravaux, Pasquale Nardelli, Jeffrey P. Jacobs, Fabio Guarracino, Nicoló Patroniti, Bas C. T. van Bussel, Iwan C. C. van der Horst, Fabio Silvio Taccone, Silver Heinsar, Kiran Shekar, Michael Yamashita, Nchafatso G. Obonyo, Anna L. Ciullo, Jordi Riera, Heidi Dalton, Anson Wang, Akram M. Zaaqoq, Graeme MacLaren, Kollengode Ramanathan, Jacky Y. Suen, Gianluigi Li Bassi, Kei Sato, John F. Fraser, Giles J. Peek, Rakesh C. Arora and on behalf of the COVID-19, Critical Care Consortium (CCCC) Cardio/ECMOCard in Perfusion.