Intraprocedural Management of Pulmonary Embolism Patients Undergoing Transcatheter Treatment, and Ventilatory and Cardiac Support

Acute pulmonary embolism (PE) is a life-threatening condition that often requires rapid, multidisciplinary intervention, especially in patients at high risk for early death. The development of transcatheter treatment methods, such as catheter-directed thrombolysis and mechanical thrombectomy, has significantly expanded treatment options for these patients, especially when systemic thrombolysis is contraindicated or ineffective.¹ During transcatheter procedures for PE, vigilant intraprocedural management is crucial, as abrupt haemodynamic or respiratory compromise may occur. In this context, robust respiratory and cardiac support measures, such as non-invasive and invasive ventilation strategies and temporary mechanical circulatory support (MCS), play a pivotal role in stabilising patients during high-risk (HR) interventions.² The judicious application of these supports is critical to optimise right ventricular (RV) function and systemic oxygenation throughout the procedure and in the post-procedure period.

Successful intraprocedural care requires a coordinated approach that anticipates sudden decompensation and incorporates both advanced supportive technologies and catheter-based intervention. This paper reviews key principles of ventilatory and cardiac support in the intraprocedural management of PE patients, with an emphasis on maintaining haemodynamic and gas exchange stability during transcatheter treatment.

Right Ventricular Failure Development in Pulmonary Embolism

An acute PE results in a sudden elevation of RV afterload, attributable both to direct increases in pulmonary vascular resistance (PVR) due to clot burden and to neurohormonal and hypoxia-mediated feedback mechanisms.^1,3 Various molecules, including serotonin, thromboxane and histamine, likely facilitate this effect. The resulting rise in PVR leads to an overload of RV pressure and volume, which in turn reduces stroke volume (SV). To compensate, a neurohormonal cascade enhances both chronotropy and inotropy. However, the RV cannot significantly increase SV, making tachycardia the primary physiological response for sustaining cardiac output (CO). The dilatation of the RV-free wall worsens contractility as the chamber exceeds its inflection point on the Frank-Starling curve.

Meanwhile, the leftward shift of the interventricular septum hampers left ventricular (LV) diastolic filling and diminishes LV SV. As LV output falls, systemic and coronary perfusion pressures decline as well. This decrease in the coronary perfusion gradient, combined with inflammatory changes and heightened myocardial demand, triggers RV ischaemia. Consequently, a feedback loop is established where RV dysfunction leads to LV dysfunction, which further damages the RV, resulting in the so-called RV death spiral.¹

Supportive treatment for patients with intermediate-high (IHR) and HR PE undergoing transcatheter treatment, especially in cases of hypoxic respiratory failure and normotensive or obstructive shock, is crucial for comprehensive care management.

Oxygen Therapy

Respiratory failure commonly associated with PE is characterised by hypoxaemia, defined as pulmonary artery oxygen <60 mmHg alongside normal or low partial pressure of arterial carbon dioxide levels. This hypoxaemia primarily results from a mismatch between ventilation and perfusion due to mechanical obstruction or pulmonary arterial vasoconstriction. The mechanisms contributing to pulmonary vasoconstriction include the activation of thromboxanes, leukotrienes, platelet-activating factor, serotonin, endothelin and decreased nitric oxide availability.⁴ Significantly, the abnormalities in gas exchange associated with PE are influenced by the size, characteristics and extent of the embolic occlusion, as well as the underlying cardiopulmonary status, the time elapsed since embolisation and the presence of veno-arterial shunting, such as a patent foramen ovale. Oxygen therapy has been shown to reduce RV afterload and lower its mechanical work.⁵ In patients with PE, a peripheral arterial oxygen saturation (SaO₂) of 92–96% should be achieved.⁶

Supplemental oxygen is indicated when SaO₂ falls below 90%. It can be delivered using a low-flow nasal cannula, standard face masks, or non-rebreather masks. If this fails, increased respiratory support may include a high-flow nasal cannula (HFNC) and mechanical ventilation (non-invasive or invasive as needed).⁷

Patients requiring >6 l/min of oxygen should be placed on a face mask or non-rebreather mask.^1,8 Nevertheless, wearing a face mask can be uncomfortable for patients, particularly during feeding, as removing the mask may influence SaO₂ levels and increase the demand for supplemental oxygen. Moreover, the use of non-invasive ventilation increases intrathoracic pressure, which may decrease RV SV and arterial pressure, potentially leading to its failure and causing periprocedural haemodynamic collapse.⁹

High-flow Nasal Cannula

The HFNC delivers a heated (31–37°C) and humidified (44 mg/l) air–oxygen mixture with an inspired oxygen fraction ranging from 21% to 100% at flow rates of 2–80 l/min, effectively reducing intubation rates and mortality for patients with acute hypoxaemic respiratory failure.¹⁰ Additionally, HFNC reduces breathing effort and respiratory rate while improving pulmonary compliance and end-expiratory lung volume, resulting in an increase in end-expiratory lung volume and prevention of alveolar closure.^7,10

HFNC oxygen therapy is better tolerated than non-invasive ventilation, allowing patients to speak and eat without compromising their oxygen needs. It delivers a high fraction of inspired oxygen and minimal essential positive end-expiratory pressure through nasal prongs, which helps wash out dead space ventilation in the upper respiratory airways.⁷ Recent minor observational studies indicated significant improvement in respiratory distress within just 1 hour for PE patients receiving HFNC, as measured by SaO₂ and respiratory rate.^9,11 A small, randomised trial on IHR and HR PE patients with acute hypoxaemic respiratory failure showed that SaO₂ and pulmonary artery oxygen values increased more rapidly and for a longer duration with HFNC compared to nasal cannula. In serial measurements of fingertip SaO₂, the first significant difference was detected between baseline and 30 minutes, and the second change after 30 minutes was observed at 24 hours. In the HFNC group, the respiratory rate was significantly lower at 30 minutes compared to baseline and this significant decrease continued at 4 and 24 hours. In the nasal cannula group, the respiratory rate was reduced at 30 minutes compared to baseline, but not at later time points.¹²

HFNC, in contrast to conventional oxygen therapy, decreases the need for increased respiratory support, alleviates dyspnoea and boosts patient comfort.¹³ Further, HFNC oxygen therapy reduces complications often associated with non-invasive ventilation, such as nasal trauma, skin irritation, pressure ulcers, air leaks and discomfort caused by high-pressure delivery.¹⁴ It is crucial to note that HFNC has no reported harmful haemodynamic effects.

When transitioning from standard oxygen therapy to HFNC, initial settings typically include a flow rate of 50–60 l/min, 100% fraction of inspired oxygen and adjustments made to reach the target SaO₂ of 92–96%. The starting temperature is set at 37°C and can be adjusted down to 34°C if necessary for improved tolerance.¹⁵ Proper cannula positioning is crucial; the prongs must fit snugly in the nostrils while avoiding complete blockage. Oxygen levels are adjusted to maintain adequate saturation and the flow rate is customised to the patient’s inspiratory needs. After beginning therapy, it is crucial to monitor the patient’s clinical status and perform arterial blood gas tests within the first hour. Signs of respiratory failure, including increased work of breathing, worsening gas exchange and tachypnoea, may indicate a failure of HFNC and suggest the need to escalate to non-invasive ventilation or intubation, as appropriate.¹⁶

The major limitation of current HFNC devices is the difficulty of patient transportation. Intrahospital transfer of patients with PE in critical condition is frequent, occurring between the accident and emergency department or medical wards and the catheterisation unit or for diagnostic tests. During these transfers, oxygenation should be maintained. Most of the previous HFNC systems did not feature an integrated power supply, relying instead on a ‘pluggable’ external power source when switching between clinical units. However, the latest generations of HFNC devices now include a battery that offers up to 1 hour of operation, facilitating safe patient transport.^15,16

Non-invasive Positive Pressure Ventilation

Research supporting non-invasive positive pressure ventilation (NPPV), specifically bilevel positive airway pressure, for patients with PE is scarce.^17,18 NPPV can help reopen collapsed lung areas, counteract pulmonary vascular shunting and provide time for the physician to initiate resuscitation.^17,19 However, positive pressure can also adversely affect RV function and increase right atrial pressure and reduce preload, potentially worsening haemodynamic instability.¹⁸ For borderline IHR PE patients in normotensive shock or HR PE patients, even minor changes in volume status can markedly impact CO due to their vulnerability and the already compromised RV preload, when exposed to positive pressure.¹⁹ Some data indicated a notable reduction in preload with NPPV application when the mean right atrial pressure was <10 mmHg.¹⁹

HFNC provides a lower level of positive airway pressure compared to NPPV, resulting in lower mean airway pressures and a reduced risk of impairing venous return. Therefore, it might be a safer choice for oxygen delivery in haemodynamically unstable PE patients, while still helping to reduce the work of breathing and enhance oxygenation.²⁰

Nevertheless, NPPV can be beneficial when used cautiously, with low pressures, close haemodynamic monitoring and readiness to escalate care if decompensation happens. Since NPPV can be stopped quickly if hypotension occurs, it remains a viable option for patients struggling with oxygenation or ventilation despite supplemental oxygen support.

Invasive Mechanical Ventilation

Intubation and invasive mechanical ventilation (IMV) should be reserved for patients who do not respond to or cannot tolerate non-invasive ventilation. This approach is essential because inducing anaesthesia and positive-pressure ventilation may exacerbate cardiorespiratory instability.¹ Criteria for intubation in acute PE are not well defined. Ventilation/perfusion mismatch, hypoxaemia, bronchoconstriction and fatigue may lead to respiratory failure, but studies on patient selection are lacking.^21,22 Research on death, haemodynamic collapse, cardiac arrest, or worsening hypoxaemia in patients intubated during acute PE diagnosis is also limited. Studies have shown that even 19% of HR PE patients experience cardiac arrest during induction of intubation and another 17% suffer cardiac arrest shortly afterwards.^23,24 During induction, it is crucial to prevent additional hypoxaemia and hypercarbia, as both conditions can lead to pulmonary vasoconstriction and increased RV afterload. This can be achieved by maximising preoxygenation and consideration should also be given to the placement of the arterial line. The IMV, by creating positive intrathoracic pressure, can reduce venous return and worsen RV dysfunction, leading to decreased LV preload and CO in patients with PE and obstructive shock.²⁵ Thus, it is crucial to establish adequate venous access via either a central line or peripheral lines and ensure adequate vasopressor dosage to maintain mean arterial pressure and safeguard RV perfusion pressure.^1,26 Anaesthetic drugs that are more likely to induce hypotension, such as propofol, should be avoided. In such cases, etomidate is preferred for induction, or ketamine can be used unless contraindicated.^1,27

Positive end-expiratory pressure should be used cautiously and, if feasible, targeted at 0 cmH₂O.²⁸ Expert recommendations suggest using low tidal volumes, approximately 6 ml/kg of lean body weight, to keep the end-inspiratory plateau pressure under 30 cmH₂O.^1,28 However, it should be emphasised that IMV alone often cannot correct hypoxaemia without simultaneous pulmonary reperfusion.²⁹ IMV in HR PE patients is linked to a 40–80% higher risk of death compared to those not ventilated.²¹ Additionally, among PE patients undergoing surgical embolectomy under anaesthesia, those who required preoperative IMV tend to have poorer outcomes.²¹ Considering all aspects of intubation and IMV initiation, continuous haemodynamic monitoring (including invasive arterial and central venous measurements), echocardiographic guidance and having extracorporeal support on standby in severe cases of collapse are essential. Based on a small study performed on HR PE patients treated with systemic thrombolysis hospitalised in the intensive care unit, analysis of end-tidal carbon dioxide tension appears to be useful for non-invasive monitoring and efficacy of thrombolysis.³⁰

A recent prospective FLAME study, which was multicentre, nonrandomised and observational, revealed that up to 25% of all HR PE patients required IMV upon their hospital admission, whereas 5.7% of patients with HR PE treated with percutaneous thrombectomy with the large-bore FlowTriever system (Inari Medical) required intubation during or shortly after the procedure.³¹ Other studies conducted primarily on patients with IHR PE reported a need for escalation to IMV in less than 1% of cases.^32–38 Tables 1 and 2 summarise key risk factors and mitigation strategies in respiratory support for PE patients.

Fluid Therapy

IV fluids must be administered carefully, because excessive volume expansion can impair RV function by causing mechanical overstretching, a leftward shift of the interventricular septum, or triggering reflexes that decrease contractility, ultimately lowering systemic CO.^1,26,28 As a result, determining the appropriate amount of fluid to administer can be particularly difficult. Current guidelines recommend a moderate fluid challenge (≤500 ml of saline or Ringer’s lactate) for patients with low central venous pressure (CVP), as it may improve cardiac index (CI) in those with PE.¹ However, this recommendation is based on findings from a small study conducted nearly 25 years ago, which showed an increase in CI from 1.7 to 2.1 l/min/m² after the infusion of 500 ml of dextran over 15 minutes in normotensive patients with acute PE and a low CI.^1,39

These guidelines emphasise the necessity of using CVP monitoring (e.g. through ultrasound imaging of the inferior vena cava [IVC]) to guide volume loading. If signs of elevated CVP arise, further volume administration should be avoided.¹ In practice, a reliable, well-validated method for predicting volume responsiveness in acute RV dysfunction is not available. Thus, clinical judgement is crucial, requiring an individualised assessment for each patient.

Proper and adequate fluid therapy is particularly important for patients undergoing transcatheter procedures, especially aspiration thrombectomy, where there is a risk of blood loss that may increase the likelihood of haemodynamic instability in patients with RV failure. Therefore, ensuring appropriate vascular bed filling is crucial for these patients and requires close attention during the perioperative period.

Right Ventricle Pharmacological Support

Haemodynamics may require support with vasopressors and sometimes inotropes while awaiting definitive treatment. Occasionally, vasoactive medications address residual RV stunning after clot debulking. Vasopressors maintain systemic and coronary perfusion to the pressure-overloaded RV. The goal is to increase the systemic vascular resistance (SVR) to PVR ratio. Peripheral vasopressor administration seems to be safe in the first 24 hours.²⁶ Nonetheless, central venous access is preferred, especially for patients who require more aggressive treatment, including higher vasopressor doses, the administration of more than one vasopressor, fluid resuscitation, additional IV medications, or blood product transfusion.⁴⁰

Based on our experience and considering that the placement of the central venous catheter is time-consuming and may delay the prompt initiation of vasopressor therapy, the authors of this review suggest inserting the central venous catheter before transcatheter intervention in patients with IHR PE and impending haemodynamic instability (normotensive shock), especially if a less-experienced operator performs the procedure.

Vasopressors

Noradrenaline is beneficial because α₁-mediated vasoconstriction increases blood pressure and enhances venous return. Additionally, β₁ stimulation strengthens both the right and left ventricles and is the preferred vasopressor in cases of cardiogenic shock.⁴¹

Noradrenaline is the preferred initial vasopressor for PE at a dose of 0.05–1.0 μg/kg/min because it maintains coronary perfusion pressure and improves SVR, without increasing pulmonary PVR.⁴² It should be adjusted to maintain the mean arterial pressure above 65 mmHg.⁴³ Higher doses of noradrenaline can theoretically increase PVR, but this seldom has a practical clinical impact.²⁶ In clinical trials, a threshold of about 0.5 μg/kg/min of noradrenaline is commonly regarded as a marker for refractory shock.⁴⁴ Phenylephrine, an α-adrenergic receptor stimulant, should be avoided because it only increases systemic afterload.²⁶

Similarly, vasopressin, a non-catecholamine vasoconstrictor, is advantageous since it induces selective peripheral vasoconstriction without elevating PVR.⁴⁵ Another possible benefit of vasopressin is its significant effect on constricting renal efferent arterioles. This action boosts renal filtration and urine production, decreasing blood volume.⁴⁶ Considering the potentially detrimental impact of excessive intravascular volume in PE, this might provide another rationale for the haemodynamic effects of vasopressin on the pulmonary circulation. Despite limited data on vasopressin use in PE, it is recommended to add it as a second vasopressor at a dose of 0.01–0.03 U/min when the noradrenaline dosage exceeds 15 μg/min or if there is a subpar response to noradrenaline.^47,48

Inotropes

Inotropes support RV contractility, thus enhancing LV filling. The objective is to enhance CO for optimal organ perfusion. Among the available inotropic agents, dobutamine is the most frequently employed in instances of PE. Dobutamine has demonstrated its ability to reduce pulmonary arterial elastance and resistance while also restoring RV–pulmonary artery coupling in cases of RV failure due to pressure overload.⁴⁷ Although European Society of Cardiology guidelines indicate that dobutamine dosage for patients with PE and low CI, with normal blood pressure, should range between 2 and 20 μg/kg/min, raising the CI beyond normal levels could exacerbate ventilation/perfusion mismatch by further redistributing blood flow from partially obstructed vessels to unobstructed.^1,28 Conversely, higher doses may lead to tachycardia and elevated oxygen consumption.⁴⁵ Administering low doses (2–5 μg/kg/min) increases CO while simultaneously reducing PVR.

Dopamine influences CO and vascular tone depending on the dosage. Its use has declined due to unpredictability, a higher risk of arrhythmias and the availability of superior alternative medications.⁴⁷

Levosimendan is an innovative inodilator that functions through three primary mechanisms: enhancing the sensitivity of cardiomyocyte troponin C to calcium, promoting vasodilation by opening potassium channels; lowering intracellular free calcium, inhibiting phosphodiesterase 3; and providing cardioprotective effects.⁴⁹ Initial research indicated that levosimendan in PE enhances RV inotropy and reduces PVR, primarily due to its capacity to induce pulmonary vasodilation.^5,50

Phosphodiesterase-3 inhibitors (milrinone or enoximone) enhance the inotropic effect by decreasing the breakdown of cyclic adenosine monophosphate within cells, which in turn increases the intracellular concentration of calcium. Phosphodiesterase-3 receptors are absent in the pulmonary vasculature. Thus, phosphodiesterase-3 inhibitors exert a positive inotropic effect on the RV without the deleterious effects on PVR that occur with catecholamines.⁵¹ However, there is a lack of sufficient scientific evidence and no official guidelines supporting a specific protocol for administering phosphodiesterase-3 inhibitors in severe forms of PE.

Phosphodiesterase-3 inhibitors and levosimendan require a longer time (hours) to achieve their maximal effect and have a prolonged duration of effect lasting hours to days. In contrast, dobutamine offers the advantage of a rapid onset of effect within minutes, but it increases myocardial oxygen demand, which can precipitate myocardial ischaemia.²⁹

In cases of RV dysfunction, the administration of inotropes should ideally be confirmed by assessing inadequate CO, even after blood pressure has been normalised. Due to possible adverse effects, inotropes should only be employed in situations of low CO and discontinued as soon as possible. Table 3 summarises the haemodynamic effects and dosages of pharmacological agents used in the treatment of RV failure.

Right Ventricle Mechanical Circulatory Support

MCS serves as a bridging therapy, helping to maintain sufficient blood flow to the body and allowing time to correct the underlying pathology. Whereas medical therapy can moderately improve afterload and contractility, MCS systems offer more effective CO for critically ill patients. Percutaneous temporary right-ventricular assist devices enable prompt support without surgical intervention. These devices are designed for shorter usage periods (typically <14 days). The mechanism of action classifies them into direct RV bypass systems, such as the Impella RP (Abiomed) and the ProtekDuo (LivaNova) system, and indirect RV bypass systems, including veno-arterial extracorporeal membrane oxygenation (VA-ECMO). Rarely can other pumps be configured with cannulas placed in the right atrium and pulmonary artery to support the RV. MCS should be considered in HR PE with cardiac arrest, refractory shock and/or contraindications to or failure of systemic thrombolysis.⁵² As MCS systems are large-bore devices, percutaneous cannulation should be performed under ultrasound guidance to minimise access site bleeding.

Impella RP

The Impella RP is a 23 Fr axial flow pump inserted percutaneously through the femoral vein. It moves blood from the right atrium into the pulmonary vasculature. The motor is entirely internal, with only the driveline exiting through the femoral vein. The device delivers up to 4 l/min of blood flow. The main disadvantages are the inability to incorporate an oxygenator if hypoxia develops and the limitation to femoral access, as it cannot be placed via the jugular vein.

Published data on the use of the Impella RP for PE are limited to small case series. However, these have shown haemodynamic improvements when used alongside definitive PE treatments.⁵³ Impella RP seems suitable for patients with persistent RV failure without hypoxaemia after successful clot reduction.^54,55 Its effectiveness in patients with a large thrombus burden remains uncertain and high-flow pumping into obstructed pulmonary arteries should rather be avoided.

It should be emphasised that the Impella RP might be considered for patients with ongoing RV failure who have previously undergone clot reduction; otherwise, the pump may push blood against a persistent blockage in the left pulmonary artery.

ProtekDuo

The ProtekDuo is a 29 Fr or 31 Fr dual-lumen catheter accessed through the internal jugular vein and positioned in the pulmonary artery. Blood is drawn from the right atrium via the inflow cannula and routed through an external pump, which then delivers it back to the pulmonary artery through the outflow lumen. Because the pump is external, an oxygenator can be easily added before the blood is returned to the pulmonary artery.⁵⁶ Like the Impella RP, the ProtekDuo is most effective in patients with persistent RV failure after significant clot reduction in the pulmonary arteries.⁵⁷ While small case series have been published, large-scale studies evaluating its use specifically in PE are lacking.^56–58

Veno-arterial Extracorporeal Membrane Oxygenation

VA-ECMO is the most widely used MCS for patients with HR PE. However, rates of use are only 0.2% of all PE.⁵² ECMO draws blood from the jugular vein and/or IVC, oxygenates it in an external circuit and then returns it to the body through the femoral artery into the aorta. VA-ECMO allows stabilisation without the immediate need for definitive clot debulking by fully bypassing the heart and lungs.

VA-ECMO use has expanded significantly worldwide with increased availability, though many centres still face challenges with rapid initiation. Notably, with more mobile ECMO circuits now available, cannulation can be performed anywhere in the hospital, allowing teams to assist other hospitals that lack in-house capabilities.⁵⁹

Veno-venous ECMO is not recommended for acute PE, as it returns blood to the venous system, failing to reduce RV preload or distension.²⁶

While VA-ECMO provides full cardiopulmonary support, it requires large-bore arterial access and full-dose anticoagulation (targeted activated clotting time of 180–220 seconds or activated partial thromboplastin time 60–80 seconds), which carries higher risks of complications such as bleeding and distal limb ischaemia.⁶⁰ These risks are particularly pronounced when VA-ECMO is initiated shortly after systemic thrombolysis due to the heightened bleeding risk associated with large-bore arterial cannulation. A need for a large-bore arterial cannula is the main limitation as compared to percutaneous RV assist device support. A report from a large registry assessing outcomes in patients requiring VA-ECMO revealed that the bleeding rates at the cannulation site were 15%, with around 7% experiencing limb ischaemia.⁶¹ Moreover, bleeding rates post-thrombolysis can reach up to 61–100%.^60,62 Of note, a single-centre study showed similar survival to discharge between ECMO patients who received systemic thrombolysis and those who did not, despite significantly higher major bleeding events.⁶³ The optimal timing for safe cannulation following thrombolysis is not clearly established. Nonetheless, in cases of progressive shock or continuous cardiac arrest, VA-ECMO frequently becomes the sole option for possible survival.

In cases of profound shock or cardiac arrest, VA-ECMO is preferred as initial support before administering systemic thrombolysis. Once stabilised with VA-ECMO, patients can proceed to more definitive treatment for their PE, percutaneous, or, in rare cases, surgical pulmonary embolectomy or recovery.⁶⁴ An extensive European report on VA-ECMO in PE showed a 30-day survival rate of 51.7%, the highest at 85.7% for surgical embolectomy patients.⁶⁵ However, no patients received catheter-based therapy. In a high-volume centre, early initiation in selected HR PE patients led to a 14.9% mortality rate.⁶⁶ VA-ECMO can also be used during CPR (extracorporeal CPR). Still, survival rates are low (20–40% for most ECMO indications) and cannulation needs to be performed within the first hour of CPR initiation.⁶⁷

Systemic thrombolysis is not absolutely contraindicated in patients on VA-ECMO; however, its administration is controversial.⁶⁸ Limited data are available on the use of catheter-directed local low-dose thrombolysis (CDL) during VA-ECMO and the potential for major bleeding complications associated with CDL raises concern.^64,69 CDL necessitates 5–7 Fr venous access and can be conducted while on full ECMO support without significantly increasing the risk of air embolism.⁷⁰

However, patients on VA-ECMO and severe RV dysfunction may require high flow rates to decompress RV, thus limiting the efficacy of the CDL itself. While CDL allows for direct thrombolytic administration to the clot, proper VA-ECMO RV decompression restricts blood flow into the pulmonary circulation. The effect of these reduced flows on the efficacy of CDL therapy remains uncertain. Some researchers suggest using CDL for patients with distal segment and subsegment clots, which are often more challenging to access via percutaneous thrombectomy.⁷⁰ This situation often arises after CPR, as chest compressions can dislodge thrombi, leading to migration further down the circulatory system. Establishing a well-defined CDL protocol and identifying the ideal flow range for this therapy alongside ECMO intervention is crucial.⁷⁰

Several considerations should be made when performing catheter-based medium- or large-bore thrombectomy in patients on VA-ECMO. Due to the size of venous access, it is essential to reduce the VA-ECMO flows when the sheath is placed into the venous system and each time the catheter is introduced through the sheath to prevent air from entering. When performing percutaneous thrombectomy in conjunction with VA-ECMO, the perfusionist should turn the flow rate down as low as tolerated each time the catheter is inserted. Once the catheter is positioned, normal ECMO flow can be restored while the thrombectomy is being performed. When the catheter is withdrawn or aspirated, blood is returned to the patient and ECMO flow rates must again be reduced to avoid potential complications related to air entrapment and embolisation.⁷¹

Operators need not worry about the IVC accommodating the ECMO venous cannula and the large-bore sheath for thrombectomy. The IVC often dilates and remains compliant in patients experiencing RV failure, which aids the placement of both devices via femoral access. In smaller patients, operators might consider using both the jugular and femoral veins to facilitate VA-ECMO from one site and the thrombectomy catheter from the other.

VA-ECMO support is generally maintained for 3–5 days, regulated by haemodynamics and RV function. To assess these parameters, patients undergo a bedside weaning trial that involves lowering VA-ECMO flows. If transthoracic echocardiography indicates normal right and left ventricular function and the patient can sustain a CI >2.0, VA-ECMO support is discontinued. If there is evidence of brain death or if complications are considered non-survivable, transitioning to comfort care is appropriate.⁶³

Device Selection

There are no direct comparative studies between support devices for acute PE. Device selection typically depends on the patient’s condition, institutional resources and expertise. The Impella RP and ProtekDuo systems are relatively quick to deploy and do not require a perfusionist. VA-ECMO, while more labour-intensive and requiring a perfusionist, provides more comprehensive cardiopulmonary support. Table 4 summarises the configuration, haemodynamic effects and device-specific selection considerations of different MCS systems.

Sequencing Interventions

The optimal timing and sequence of advanced interventions such as VA-ECMO, mechanical thrombectomy and vasopressor or inotropic support in PE remain areas of significant clinical uncertainty and ongoing research. Current evidence indicates that, for HR PE with refractory haemodynamic instability or cardiac arrest, immediate initiation of vasopressor/inotropic support is generally the first step to stabilise circulation while definitive therapies are arranged.¹ VA-ECMO may serve as a life-saving bridge to both reperfusion and recovery in patients unresponsive to initial resuscitation, especially when thrombolysis is contraindicated or unsuccessful. Yet, VA-ECMO is most effective when used adjunctively rather than as sole therapy.⁶⁴ Catheter-directed thrombectomy is considered for patients with HR and IHR PE who have persistent RV dysfunction or deteriorating haemodynamics despite anticoagulation, or in those for whom thrombolysis is unsuitable.⁶⁴ Early thrombectomy may improve pulmonary artery pressure and possibly outcomes compared to delayed intervention; however, no definitive mortality benefit has yet been confirmed.^45,50 The integration and optimal sequencing of these modalities – supportive therapy, VA-ECMO and advanced reperfusion –requires multidisciplinary assessment and should be tailored to individual patient profiles, institutional expertise and resource availability.⁴⁵ Nonetheless, the available evidence, drawn primarily from observational registries and single-centre series, highlights substantial variation in practice and underscores key knowledge gaps, particularly regarding patient selection, timing and the balance between bleeding and thrombotic risks. Future randomised comparative studies are urgently needed to delineate better the synergistic effects, optimal protocols and long-term outcomes of combined VA-ECMO and catheter-directed reperfusion strategies in this critically ill population. Since the clinical landscape is rapidly evolving with limited high-quality randomised data, close coordination between PE response teams and further prospective research is critical to define best practices in this area.

Conclusion

The haemodynamic and respiratory support in PE is particularly complex and heterogeneous and should be tailored to the individual patient. A multidisciplinary approach is frequently required to navigate the complex process of selecting the best treatment options for advanced PE. We summarised the main components of supportive care in PE in Figure 1. A cautious approach is recommended, as the harms may sometimes outweigh the benefits when used inappropriately. Current recommendations are primarily based on studies with low levels of evidence, most of which consist of expert opinions and observational studies. While progress has been achieved, further high-quality research is essential to enhance clinical decision-making about supportive therapy in PE. We hope this review stimulates interest in this evolving field.