The Impella (Abiomed, Danvers, MA) mechanical circulatory support (MCS) system is a catheter-based continuous flow pump that is typically placed percutaneously or by surgical cutdown into the femoral or subclavian artery. The hemodynamic effects of the Impella catheter are to improve systemic perfusion and provide ventricular unloading in the setting of high-risk percutaneous coronary intervention and in the treatment of cardiogenic shock. The Impella catheter has become a common MCS device used in medical and surgical cardiac intensive care units, and as such intensivists must have a core competency with its management. The purpose of this review is to describe how to manage, reposition, and wean patients from the Impella catheter.
There are currently five Impella catheters that provide left ventricular (LV) support (Figure 1). The structural design of each of these LV support catheters is grossly similar (Figure 2). The tip of the catheter has a flexible pigtail, intended to prevent mechanical injury of the ventricle (absent on the larger LD and 5.5 models). The pigtail attaches to a radiopaque/echogenic structure termed the “teardrop” which is contiguous with the inlet area, through which blood enters the ventricular end of the catheter’s cannula. The aortic end of the cannula houses a microaxial motor which spins an Archimedes’ screw impeller that draws blood through the cannula to the outlet area in the aortic root.
There is a pressure sensor built into the aortic end of the cannula that is used to produce a “placement signal” waveform tracing (Figure 3A). That waveform depicts the pressure gradient across the intra- and extraluminal surface of the cannula, and when the device is correctly positioned, the intra- and extraluminal pressures reflect the pressure within LV and aortic root, respectively. The morphology of the placement signal resembles that of an arterial waveform, which can be counterintuitive as the peak of the waveform occurs in diastole when the gradient between the LV and aorta is greatest, whereas the nadir occurs during systole when it is least. Catheters with “SmartAssist” have an upgraded optical pressure sensor that directly measures aortic pressure and uses changes in motor current to infer an estimated LV pressure waveform. It is important to note that this derived ventricular pressure is not an accurate measure of the true LV pressure and thus may not directly replace the value of monitoring the pulmonary arterial capillary wedge pressure via a pulmonary artery catheter.
The Impella catheter connects to a bedside controller that displays data about device performance, powers the motor, and delivers a heparinized dextrose solution (purge fluid) to the motor housing.
The smaller diameter Impella 2.5 and CP are typically inserted percutaneously under fluoroscopic guidance in the cardiac catheterization laboratory; however, transesophageal echocardiography (TEE) guided placement is also possible.1 The larger diameter Impella 5.0 and 5.5 are typically inserted by surgical cutdown via a prosthetic graft sewn onto the subclavian artery and are positioned under fluoroscopy or by TEE. The lesser used Impella LD is reserved for direct transaortic placement intraoperatively. The Impella RP is a right ventricular (RV) support system that is percutaneously positioned in the pulmonary artery via the femoral vein under fluoroscopy. As the Impella RP is not widely used, this review will focus only on the left-sided catheters; however, many of the concepts discussed are transitive.
Ultrasound Assessment of Catheter Positioning and Repositioning
The Impella catheter must be adequately positioned to provide optimal hemodynamic support while minimizing the risk of complications, including hemolysis, interference with the mitral apparatus, suction events, or provocation of ventricular arrhythmias. The catheter is not tethered to any internal structure and is prone to migration which occurs more frequently when the catheter is placed percutaneously. Catheter position should be assessed daily and in response to unexpected clinical changes or specific device alarms. Notably, catheter position is best assessed ultrasonographically, as radiography does not accurately identify the cannula position relative to cardiac anatomic structures. Additionally, a gross assessment of cannula depth can be inferred by contrasting the placement signal and LV pressure waveforms on SmartAssist capable devices (see controller alarm section.)
The most reliable and accurate transthoracic acoustic widow to assess catheter position is the parasternal long-axis window because it will provide a view of the catheter from an orthogonal angle of insonation (Figure 4). In cases where transesophageal ultrasound is used for catheter placement or repositioning, the midesophageal long-axis view (120°) is the most reliable and accurate to assess catheter depth. A multi-beat acquisition of the catheter should be obtained while panning through the LV cavity, to avoid catheter foreshortening and consequent incorrect assessment of catheter depth. The aortic annulus and the ventricular end of the cannula must be well visualized in a single image to make an accurate measurement. Optimal imaging often requires off-axis parasternal long-axis views obtained by fanning and rotating the probe until the entire length of the cannula and the aortic annulus are seen. The necessary images may be particularly difficult to obtain if the Impella device is medially or laterally oriented. We recommend the following steps to optimize imaging: minimize the depth and narrow the sector width to the target field of view, angle the ultrasound probe to achieve a horizontally oriented view of the cannula, and adjust the gain to best highlight the target structures.
The catheter depth is defined as the distance from the aortic annulus to the ventricular end of the cannula (Figure 4). The cannula appears as two bright echogenic, parallel lines, sometimes referred to as “the railroad tracks”, which ends at the inlet area. The optimal depth for the Impella 2.5, CP, 5.0, and LD is 3 cm to the beginning of the inlet area, and readjustment should be considered if the depth is more than 0.5 cm from this target. The optimal cannula depth of the Impella 5.5 is 4.5 cm +/- 0.5 cm to the beginning of the inlet area, as this model has a longer cannula. Notably, the device manufacturer suggests measuring catheter depth from the aortic annulus to the middle of the echolucent inlet area, (i.e., 0.5 cm more than the distances stated above). However, in our experience, measuring structures that can be directly visualized with ultrasound is a more practical and reproducible strategy, and thus we recommend measuring from the aortic annulus to the “end of the railroad tracks”.
An additional method to ensure proper depth is to interrogate the aortic root with color Doppler from the parasternal long-axis view. A broad mosaic color Doppler pattern caused by artifact from the motor will be seen, which should be limited to the aortic side of the valve. If a significant color Doppler signal is observed below the valve (in the absence of significant aortic regurgitation), the device is likely too deep.
Optimizing the spatial relationship between the catheter and adjacent intracardiac structures is necessary to minimize device-related complications. The cannula portion of the catheter is built with a 30° bend (except for the Impella LD, which has a straight design). The purpose of the bend is to orient the catheter toward the apical anteroseptal portion of the LV and away from the posterolateral wall, papillary muscles, and mitral apparatus. If the catheter is incorrectly oriented, the pigtail can become caught in the mitral apparatus, a segment of the catheter may restrict mitral valve opening, or mobile portions of the mitral apparatus may be drawn into the inlet area. Furthermore, crowding of the inlet or outlet areas increases shear stress on the red blood cells pumped through the cannula, accelerating the rate of hemolysis caused by the device. The size and function of the left and right ventricles as well as interventricular septal position should also be assessed, as low flow and suction alarms can be caused by over-decompression of the LV, RV failure, and/or obstructive physiology.
We recommend that repositioning of the Impella catheter be performed by two people, one to obtain real-time ultrasound images and one to manipulate the catheter. Once the imager has a nonforeshortened image of the catheter in the parasternal long-axis view, the Impella motor speed should be temporarily set to power level P2, which reduced the risk of damaging the submitral apparatus during the catheter manipulation. After making note of the catheter depth from the vascular access site, the nonimager should then loosen the vascular access site Tuohy-Borst lock (Figure 5) and rotate, advance, or withdraw the catheter as appropriate to optimize its position. Rotation can often be difficult and applying more than a full 360 degrees of torque is often necessary. Additional torque can be achieved by rotating the red Impella plug (Figure 5) at the proximal end of the catheter in the desired direction. If the catheter pigtail is hooked on the mitral apparatus and/or papillary muscle, it may be necessary to first advance the catheter deeper into the ventricle and then rotate the catheter to disengage it from the valvular structures. It is important to recognize that adjustments from the vascular access site are not necessarily transmitted to the cannula in a 1:1 fashion due to slack and/or torque that may exist or be introduced into the catheter. For this reason, we recommend a conservative approach to catheter manipulation with a “the enemy of good is perfect” philosophy. After advancement, always remove any slack by slowly pulling back on the catheter until cannula movement is observed. After any adjustment, return the power level back to the desired setting and then reassess catheter depth, orientation, and mitral valve function before tightening the Tuohy-Borst lock and making note of the final vascular access site depth.
There are two indications for anticoagulation when using the Impella catheter. The first is to prevent clot formation on the motor, a potentially catastrophic event that is avoided by delivering a heparinized dextrose purge solution to the motor which creates a liquid interface between the motor housing and the patient’s blood. The second indication is to prevent clot formation on the catheter itself and potential embolization into the patient. Infusion of purge solution alone infrequently results in the desired therapeutic systemic anticoagulation and an additional infusion of parenteral heparin is often required. Standard therapeutic anticoagulation targets are; an activated clotting time of 160–180 sec, a heparin antifactor Xa of 0.3-0.5, and/or activated partial thromboplastin time of 60–90 sec. In cases of heparin-induced thrombocytopenia, use of argatroban or bivalirudin in place of heparin in the purge solution has been reported to be safe and effective.2,3 Notably, a rising purge pressure may reflect thrombus formation in or around the motor.
Shearing of red blood cells is a common and clinically relevant problem with the Impella catheter. While a small amount of hemolysis is unavoidable, significant hemolysis can quickly cause pigment nephropathy and further complicate the management of an already critically ill patient. To monitor the severity of hemolysis we recommend daily monitoring of serum creatinine, and plasma-free hemoglobin (PFH) or lactate dehydrogenase (LDH). A PFH >40 mg/dL or acute increase in either PFH or LDH suggests increasing hemolysis and warrants intervention.4 The main causes of Impella-related hemolysis are crowding or partial obstruction of the inlet or outlet areas due to poor positioning, thrombus formation in the cannula, operating at the higher range of the power settings, and subtherapeutic anticoagulation.
To understand the hemodynamics of a patient in cardiogenic shock receiving Impella support, a pulmonary arterial catheter is recommended.5 The overall weaning strategy is to achieve adequate organ perfusion at the lowest device power setting to minimize device-related complications and to determine candidacy for device removal. Serial assessment of native cardiac function and organ perfusion using clinical, hemodynamic, imaging, and laboratory data should be performed as the Impella support is weaned. Our practice is to integrate clinical factors, such as mean arterial pressure, heart rate, and urine output with invasive hemodynamics, and lab data—specifically, serum lactate and pulmonary arterial oxygen saturation every 6 hours. Based upon these metrics we make changes to the device power level, inopressor dose, afterload reduction regimen, and diuresis goals as indicated to maintain sufficient but not excessive cardiac support while optimizing RV and LV filling pressures. Our typical hemodynamic targets are a mean arterial pressure of 60–70 mm Hg, a right atrial pressure of 8–12 mm Hg, a pulmonary arterial wedge pressure <15 mm Hg, and a cardiac index >2.0 L/min/m2. If the data suggests that the patient is over-supported, the Impella is weaned by one or two power levels. In general, if the patient subsequently develops oliguria, tachycardia, lactate >2 mg/dL, or a cardiac index <2.0 L/min/m2 we will resume the prior level of cardiac support provided by the Impella. Once perfusion goals are met and proven to be stable at power level P2, the Impella catheter should be removed.
It is not uncommon that some patients with severe cardiac dysfunction do not readily demonstrate the ability to wean from the Impella. The question that then arises is when to favor escalating inotropes to assist in device weaning. In these situations, one must weigh the risk of time-dependent device-related complications with inotrope-related increases in myocardial oxygen demands. While there is little data to draw from, it follows that inotropes should be minimized or avoided in patients with unrevascularized coronary disease or active ischemia. Further to this point, cases of Impella weaning intolerance or clinical decompensation after explantation beg the complicated question of when to consider escalating inotropes and/or pursuing additional MCS as bridge-to-recovery or bridge-to-LVAD/transplant versus palliation, and must be determined on an individual basis.
Patients in cardiogenic shock supported by venoarterial extracorporeal membrane oxygenation (VA-ECMO) often require an LV unloading strategy to prevent the development of pulmonary edema, thrombus formation in the LV, and reduce LV wall stress.6 The concurrent use of VA-ECMO with the Impella catheter as an unloading strategy (ECPELLA) has become a popular MCS configuration associated with improved cardiogenic shock outcomes over VA-ECMO alone.7 The management of the Impella catheter while in the ECPELLA configuration is no different than described elsewhere in this review. With respect to weaning order of operations, we generally favor first weaning and decannulation from VA-ECMO (if possible from a pulmonary support perspective), which is based on the higher MCS complication rates and patient immobility associated with VA-ECMO cannulation.
Right ventricular failure
Like all LV assist devices, the Impella can only pump as much blood as is available to it. In cases of RV failure, Impella flows can be limited by poor RV output as well as by RV distention that shifts the interventricular septum toward the LV, which can precipitate suction events. Our practice of monitoring and managing RV function relies heavily on invasive hemodynamics and ultrasound imaging. Hemodynamically, we typically titrate fluid balance goals and inotropes to target a right atrial pressure of 8–12 mm Hg and a pulmonary artery pulsatility index >1. On imaging, if the LV appears overly decompressed due to a significant interventricular septal shift, then reducing the Impella speed, escalating inotropy to support the RV, and increasing volume removal is commonly the best course of action. However, in severe cases of RV failure or when pre-capillary pulmonary hypertension is a contributing factor, right-sided MCS and pulmonary vasodilator titration may be required, respectively.
Device-related complications occur more frequently with a longer duration of support. The most common Impella-related complications reported are hemolysis, embolic stroke, limb ischemia, access site bleeding, device migration, device malfunction, motor thrombosis, ventricular arrhythmia, and mitral valve disruption.8 Most of these complications are directly related to catheter position and anticoagulation, stressing the importance of frequent clinical assessment. In our experience, hemolysis (45%), device-related ventricular tachycardia (18%), and limb ischemia (16%) were the most common complications, and Impella repositioning was required in 26% of cases (Table 1.)
Table 1. -
Institutional Intensive Care Unit Impella Experience 2017–2021
||n = 91
| Direct aortic
|Days of Impella support (avg)
|Device repositioning required
| Ventricular tachycardia
| Limb ischemia
| Cerebral vascular accident
| Mitral dysfunction
Values are n (%) or mean.
The MedStar Washington Hospital Center institutional review board approved the electronic medical record extraction and publication of this data.
AMICS, Acute myocardial infarction cardiogenic shock. ECPELLA, venoarterial extracorporeal membrane oxygenation with the Impella catheter as an unloading strategy.
Suboptimal Impella flow and suction events can be caused by anything that reduces blood flow to the device, which is most commonly due to hypovolemia, RV failure, ventricular arrhythmias, and obstructive physiology (tamponade, pulmonary embolism, etc.). In the event of cardiac arrest, standard life-saving procedures should be followed with the caveat that the Impella should be set to power level P2 to prevent a continuous suction event. Additionally, cardiac resuscitation can cause device migration and as such, post-arrest ultrasound confirmation of position is recommended.
The Impella controller will alarm when it determines that the device may be mispositioned or dysfunctional. With correct positioning and function, the placement signal and motor current are pulsatile, reflecting the dynamic pressure gradient between the aorta and LV, as well as the cyclical variation in energy required to maintain the desired motor speed over the cardiac cycle (Figure 3A). Conversely, lack of expected pulsatility in the placement signal and/or in the motor current signal may reflect a problem. If both the placement signal and motor current waveforms have minimal variability, the “Impella Position Wrong” alarm will display, indicating that both the inlet and outlet areas may be on the same side of the aortic valve (Figure 3B). This should prompt urgent ultrasound assessment as the device may need to be either retracted or advanced. If the catheter is completely out of the ventricle, it should be repositioned across the valve over a guidewire. Catheters with SmartAssist, however, will more specifically identify the nature of the mispositioning and can distinguish between the ventricular and aortic placement of the inlet and outlet areas (Figure 3C and 3D). Notably, low native heart pulsatility may similarly trigger either the “Impella Position Wrong” or the “Impella Position Unknown” alarm, as the software cannot interpret the dampened amplitude of the placement signal and motor current (Figure 3E). The “Suction” alarm is triggered when sudden decreases in the placement signal pressure occur in association with lower than expected flows. (Figure 3F). Suction events may be caused by inadequate LV filling or incorrect Impella positioning with inlet area obstruction by a cardiac structure.
Important contraindications to use of the Impella MCS system are the presence of: moderate to severe aortic regurgitation, mechanical aortic valve, aortic dissection, LV thrombus, or ventricular sepal defect. While the inability to provide anticoagulation is a contraindication, there is ongoing research into nonanticoagulant purge solution alternatives.
The Impella MCS system is a relatively new technology that has become widely used for the treatment of cardiogenic shock in medical and surgical cardiac intensive care units. As with all current forms of MCS, device-related complications remain a major concern, many of which can be mitigated by adhering to a few fundamental concepts in device management.
1. Crowley J, Cronin B, Essandoh M, D’Alessandro D, Shelton K, Dalia AA: Transesophageal echocardiography for impella placement and management. J Cardiothorac Vasc Anesth. 33: 2663–2668, 2019.
2. Reed BN, DiDomenico RJ, Allender JE, et al.: Survey of anticoagulation practices with the impella percutaneous ventricular assist device at high-volume centers. J Interv Cardiol. 2019: 3791307, 2019.
3. Szymanski TW, Weeks PA, Lee Y, et al.: Anticoagulation of impella with a bivalirudin purge solution. ASAIO J. 66: e117–e120, 2020.
4. Esposito ML, Morine KJ, Annamalai SK, et al.: Increased plasma-free hemoglobin levels identify hemolysis in patients with cardiogenic shock and a trans valvular micro-axial flow pump. Artif Organs. 43: 125–131, 2019.
5. Garan AR, Kanwar M, Thayer KL, et al.: Complete hemodynamic profiling with pulmonary artery catheters in cardiogenic shock is associated with lower in-hospital mortality. JACC Heart Fail. 8: 903–913, 2020.
6. Rao P, Khalpey Z, Smith R, Burkhoff D, Kociol RD: Venoarterial extracorporeal membrane oxygenation for cardiogenic shock and cardiac arrest. Circ Heart Fail. 11: e004905, 2018.
7. Patel SM, Lipinski J, Al-Kindi SG, et al.: Simultaneous venoarterial extracorporeal membrane oxygenation and percutaneous left ventricular decompression therapy with impella is associated with improved outcomes in refractory cardiogenic shock. ASAIO J. 65: 21–28, 2019.
8. Subramaniam AV, Barsness GW, Vallabhajosyula S, Vallabhajosyula S: Complications of temporary percutaneous mechanical circulatory support
for cardiogenic shock: an appraisal of contemporary literature. Cardiol Ther. 8: 211–228, 2019.