The Impella family of devices works by drawing blood from the left ventricle through an inlet area near the pump’s tip and expelling it into the ascending aorta, providing up to several liters per minute of forward flow while simultaneously unloading the failing ventricle. The pump motor and bearing assembly sit within the blood stream, and heat and mechanical stress make that region a nidus for thrombus formation. To prevent clot from accumulating around the motor, the device continuously flushes a sterile purge solution through the narrow clearance space between the rotor and the housing; the solution then exits into the circulation. Because this purge fluid is delivered directly into the patient’s blood, whatever is dissolved in it becomes part of the patient’s systemic drug exposure. Historically, the manufacturer’s instructions have called for heparin to be added to the purge solution, so the purge line acts as a continuous low-dose heparin infusion on top of whatever systemic anticoagulation the patient is already receiving.

That arrangement is straightforward when Impella is used alone, but it becomes genuinely complicated when ECMO is added. Patients on ECPELLA support are typically managed with systemic anticoagulation to protect the oxygenator and tubing of the extracorporeal circuit, guided by protocols such as those issued by the Extracorporeal Life Support Organization, whose 2021 adult and pediatric anticoagulation guidelines consolidated contemporary practice. Layering a heparin-containing purge on top of systemic heparin, or on top of an alternative agent such as bivalirudin, creates a second, parallel route of anticoagulant delivery whose contribution to the total dose is not independently measured by standard monitoring. Clinicians therefore face a balancing act: enough anticoagulation to protect two devices and a circuit, but not so much that bleeding, the most common complication of mechanical circulatory support, becomes life-threatening. Some centers have responded by removing heparin from the purge bag entirely, running the purge with anticoagulant-free solutions such as 5% dextrose in water or sodium bicarbonate, and relying on systemic anticoagulation to protect the pump.

The question is whether an anticoagulant-free purge actually protects the device. The purge flow rate is titrated to maintain a specified pressure gradient across the purge circuit, which serves as a proxy for patency of the tiny channels that bathe the motor; if those channels clot, purge pressure rises, the purge rate climbs to compensate, and in the worst case the pump malfunctions. Prior single-center experiences, including analyses of heparin 25 units per milliliter purge solutions and separate evaluations of bicarbonate-based purge solutions in patients with cardiogenic shock supported by Impella, have suggested that non-standard purge strategies can be workable, but the data remain limited and the patients most likely to receive ECPELLA, those in profound cardiogenic shock, have been underrepresented. The new study by Nathaniel J. Martin, Christoph G. S. Nabzdyk, Scott D. Nei, Andrew N. Rosenbaum, Troy G. Seelhammer, and Patrick M. Wieruszewski was designed to examine exactly this gap: Impella device function in patients receiving combined ECMO and Impella support, stratified by whether their purge solution contained heparin or not.

The design was observational and modest in scale, and the authors are explicit about its limits. The team identified 25 patients supported with ECPELLA at their institution. Fourteen of these patients received an anticoagulant-free purge solution, either 5% dextrose or sodium bicarbonate, while the remaining 11 received a heparin-containing purge solution. The median duration of ECPELLA support was five days, a duration long enough for purge-related thrombosis to become clinically apparent if it were going to occur. The primary observation of interest was Impella device malfunction attributable to purge failure, the outcome that would most directly signal that an anticoagulant-free purge was inadequate to keep the motor housing clear. Ethical approval for the study came from the Mayo Clinic Institutional Review Board, with a waiver of informed consent, consistent with the retrospective review of clinical care.

The headline finding is stark in its simplicity. Across the entire cohort, the only episode of Impella device malfunction occurred in a patient in the heparin-containing group. No patient in the anticoagulant-free group experienced a purge-related device failure during their course of support, and device longevity appeared acceptable in both groups. Read at face value, the result suggests that dextrose- or bicarbonate-based purge solutions, paired with systemic anticoagulation, did not compromise the mechanical integrity of the Impella during a median of five days of combined support. In an environment where every additional unit of heparin delivered through the purge line potentially shifts a critically ill patient’s hemostatic balance, a strategy that removes that infusion without evident device harm is clinically meaningful, even if the evidence behind it is still thin.

The authors themselves are careful not to overstate the case, and their caution is warranted for reasons that go to the heart of how small clinical studies work. With only 25 patients and a low overall event rate, the study cannot exclude a modest increase in device malfunction risk associated with anticoagulant-free purge solutions; the single malfunction in the heparin group, far from indicting heparin, illustrates precisely how noisy such comparisons are at this scale. Bleeding outcomes, thromboembolic events, and hemolysis, all of which matter enormously in this population, are not the focus of the abstract-level reporting, and confounding by indication is a live concern: the choice of purge solution likely reflected evolving institutional practice and individual patient characteristics rather than random assignment. The conclusion the authors draw is correspondingly measured. Given the low event rate and the apparent maintenance of acceptable device longevity, they write, use of anticoagulant-free purge solutions alongside systemic anticoagulation in ECPELLA patients may be reasonable, but the strategy requires validation in an appropriately powered study.

Even so, the study contributes to a broader and increasingly visible conversation about how anticoagulation should be managed across the growing arsenal of temporary circulatory support devices. Pharmacotherapy reviews have called for optimization of anticoagulation in Impella-supported patients, noting the tension between manufacturer recommendations, center-specific protocols, and the absence of definitive trial evidence. Comparisons of bivalirudin versus heparin for systemic anticoagulation during ECMO, published in Critical Care Medicine by members of the same Mayo group, reflect the same underlying effort to rationalize hemostatic management in extracorporeal support. The purge solution sits at an unusual intersection within this landscape: it is simultaneously a device-maintenance fluid and a systemic drug delivery route, and it is frequently overlooked in discussions that focus on the more visible systemic anticoagulants. Recognizing the purge line as a pharmacologic exposure, and studying it as one, is a conceptual step forward for the field.

For the bedside intensivist, pharmacist, and perfusionist managing an ECPELLA patient tonight, the practical takeaway is not a new protocol but a new data point in an evolving conversation. Anticoagulant-free purge solutions, whether 5% dextrose or sodium bicarbonate, have now been used in a consecutive series of ECPELLA patients at a high-volume center without evident compromise of device function over a median of five days of support. That observation will not end the debate, and no one should abandon heparinized purge on the strength of 25 patients, but it legitimizes the strategy as a subject for the large, prospectively designed comparative study it deserves. Until such a study exists, the choice of purge solution will remain a matter of institutional protocol and clinical judgment, exercised within the same delicate balance that defines all of extracorporeal support: keeping the blood fluid enough to survive the machine, and the patient solid enough to survive the bleeding. This small study suggests the balance can be struck without heparin in the purge bag, and that alone makes it worth the field’s attention.

Subject of Research: Impella purge solution strategy and device function in patients receiving combined ECMO and Impella (ECPELLA) support

Article Title: Impella purge solutions in patients receiving extracorporeal membrane oxygenation

Article References: Martin, N. J., Nabzdyk, C. G. S., Nei, S. D., Rosenbaum, A. N., Seelhammer, T. G., & Wieruszewski, P. M. (2026). Impella purge solutions in patients receiving extracorporeal membrane oxygenation. Journal of Artificial Organs, 29(4), Article 64. https://doi.org/10.1007/s10047-026-01595-2

Image Credits: AI Generated

DOI: 10.1007/s10047-026-01595-2

Keywords: Impella, ECMO, ECPELLA, purge solution, anticoagulation, heparin, mechanical circulatory support, cardiogenic shock, left ventricular unloading, device malfunction, critical care, ventricular assist device

The clinical story began long before the emergency that defined it. The boy was born at full term with a normal birth weight of 2608 grams, but by two months of age he was hospitalized for poor weight gain and declining cardiac function. Oral medications stabilized him enough for discharge after a two-month stay. Genetic testing eventually revealed a nonsense mutation, designated c.153C > G, in the TAZ gene, the molecular signature of Barth syndrome. This X-linked recessive condition arises from defects in an enzyme that remodels cardiolipin, a phospholipid essential to the structure and function of mitochondria, the energy-producing structures inside cells. The result is a constellation of problems: dilated cardiomyopathy in which the heart muscle stretches and weakens, generalized muscular hypotonia, and cyclical neutropenia that leaves patients vulnerable to infection. When the child was readmitted at one year of age with severe cardiac dysfunction and a brain natriuretic peptide level of 2096 pg/mL, a marker of profound heart strain, inotropic drugs failed to rescue him and he was transferred to Osaka for transplant registration and placement of a Berlin Heart EXCOR, the only durable ventricular assist device approved for small children in Japan.

His condition deteriorated faster than the transplant system could move. On admission he weighed just 6.1 kilograms and measured 67.5 centimeters. Echocardiography showed a left ventricle dilated to 38 millimeters, a z-score of +5.7 relative to normal body size, an ejection fraction of only 15 percent, and severe mitral regurgitation; his BNP had surged to 7667 pg/mL. Soon after arrival he developed runs of non-sustained ventricular tachycardia, a dangerous rhythm disturbance, and amiodarone therapy dropped his blood pressure. The team established venoarterial extracorporeal membrane oxygenation through neck vessels as an emergency salvage measure, but acute pulmonary congestion followed, forcing the decision to implant a biventricular assist device using centrifugal pumps. For the left side, surgeons used a 6-mm apical inflow cannula and a 6-mm arterial outflow cannula; for the right side, they placed a 14-Fr inflow cannula into the inferior vena cava via the right atrium and an 8-Fr outflow cannula into the distal pulmonary artery trunk, all secured with purse-string sutures.

What happened next is the technical heart of the report. The right ventricular assist device pumped roughly 900 mL/min, yet the left-sided pump could not function efficiently. The culprits were severe pulmonary hypertension and pulmonary regurgitation, which together prevented blood pushed into the pulmonary circulation from returning effectively to the left heart. In a decisive maneuver, the surgeons relocated the right-sided outflow cannula from the pulmonary artery into the left atrium, temporarily routing oxygenated blood directly to the left side of the heart and bypassing the obstructed pulmonary circuit. Left ventricular assist device flow stabilized at approximately 600 mL/min. The strategy bought time, but it carried trade-offs: left atrial cannulation raises the risk of systemic thromboembolism, and the circuit could not be disconnected from the artificial lung that oxygenates blood outside the body. The team therefore treated the configuration as a bridge within a bridge, to be dismantled as soon as the pulmonary vasculature could tolerate normal routing.

That evaluation came quickly. Cardiac catheterization on postoperative day 5, performed during temporary interruption of right-sided pumping, recorded a pulmonary artery pressure of 21/15 mmHg with a mean of 18 mmHg and a pulmonary vascular resistance index of 3.5 Wood units times meters squared, values indicating acceptable pulmonary vascular physiology. On postoperative day 13 the outflow cannula was moved back to the pulmonary artery trunk and the artificial lung was removed from the circuit. Bedside echocardiography then showed improving right ventricular contraction, but chest X-rays revealed progressing pulmonary congestion, which the surgeons attributed to excessive pulmonary blood flow generated by full right-sided support. Their response was a careful titration: right pump flow was gradually reduced to low-flow assistance. Throughout this period the left pump maintained a stable 900 mL/min flow and central venous pressure held near 12 mmHg even when right-sided support fell to 400 mL/min or was interrupted entirely, evidence that the native right ventricle and pulmonary circulation had recovered enough to fill the left pump on their own. Concerned about circuit thrombosis during prolonged low flow, the team removed the right ventricular assist device on postoperative day 20.

The recovery of other organs lagged behind the heart but eventually followed. Acute kidney and liver failure complicated the early postoperative course, requiring continuous hemodiafiltration from postoperative day 7. Bilirubin climbed as high as 15.1 mg/dL and renal dysfunction persisted for more than a month, but hemofiltration was discontinued on day 44, total bilirubin normalized on day 61, and the child was extubated on day 55. Only after these markers of end-organ recovery did the team secure in-house approval for heart transplant registration, the prerequisite under Japanese rules for implanting the Berlin Heart EXCOR, which was converted from the temporary centrifugal left ventricular assist device 76 days after the original biventricular implantation. The boy then remained stable on the pulsatile device for 374 days before undergoing successful heart transplantation, a total mechanical support journey of roughly 450 days from the first emergency pump to the donor heart.

The Osaka team frames the case within a strategy known as bridge to candidacy, an approach better documented in adults. In Japan, pediatric candidates face an average wait of 686 days for a donor heart, even though ten-year survival after pediatric transplantation reaches 96.2 percent, and by 2022 only 68 patients under 18 had ever been transplanted in the country. The Berlin Heart EXCOR has been approved there only since 2015 and only for patients already registered for transplantation, which creates a chicken-and-egg problem for infants whose comorbidities disqualify them from listing. Adult data show the workaround can succeed: large series report that short-term mechanical support can reverse refractory cardiogenic shock long enough for conversion to durable devices, with survival comparable to direct bridge-to-transplant pathways, though right ventricular support during the conversion emerges as a mortality risk factor. Pediatric evidence is thinner, but a Berlin group’s ten-year experience with 56 children and a multicenter United States study both suggest that prior extracorporeal support does not necessarily doom Berlin Heart outcomes, while low body weight, kidney and liver dysfunction, and biventricular support do worsen survival.

The case also contributes to a re-evaluation of Barth syndrome itself as a transplant indication. Because the condition combines heart failure with skeletal myopathy and immune compromise, patients were once considered unsuitable candidates, but after the first successful transplant in 1997, London investigators reported four successful cases, and a registry analysis of 43 transplanted Barth patients found outcomes for survival, infection, malignancy, and graft vasculopathy essentially equivalent to those of other recipients. The Osaka authors add an important caveat drawn from recent metabolic studies: transplantation does not fully normalize exercise tolerance, muscle mass, or substrate metabolism, because the underlying cardiolipin defect persists in skeletal muscle and other tissues. Their patient was not systematically assessed for these parameters after transplantation, a limitation the team acknowledges, and long-term multidisciplinary follow-up of neurodevelopment and metabolic status continues.

Beyond the individual rescue, the report is a detailed argument for staged, physiology-driven management of the failing right heart in small children. When pulmonary hypertension and valve regurgitation sabotage the left-sided pump, temporarily diverting right-sided outflow to the left atrium can stabilize the circulation; once catheterization confirms that pulmonary vascular resistance has fallen, rerouting to the pulmonary artery removes the thrombotic risk of left atrial cannulation and permits artificial lung removal, and stepwise flow reduction then tests whether the native right ventricle can carry the load. In this case that sequence converted a child in refractory cardiogenic shock with failing kidneys and liver into a registered transplant candidate with a durable device. The authors suggest the approach may extend the bridge-to-candidacy strategy to other high-risk pediatric patients with biventricular failure complicated by pulmonary hypertension, a population for whom waiting lists are long and options have historically been few.

Subject of Research: Bridge-to-candidacy mechanical circulatory support and heart transplantation in a pediatric patient with Barth syndrome-associated biventricular heart failure

Article Title: Successful bridge to heart transplantation in a pediatric patient with biventricular heart failure associated with Barth syndrome: a case report

Article References: Arita, K., Kido, T., Taira, M., Watanabe, T., Narita, J., Ishida, H., Ishii, R., Ueno, T., & Miyagawa, S. (2026). Successful bridge to heart transplantation in a pediatric patient with biventricular heart failure associated with Barth syndrome: a case report. Journal of Artificial Organs, 29(4), Article 60. https://doi.org/10.1007/s10047-026-01587-2

Image Credits: AI Generated

DOI: 10.1007/s10047-026-01587-2

Keywords: Barth syndrome, heart transplantation, ventricular assist device, Berlin Heart EXCOR, biventricular heart failure, pediatric cardiology, pulmonary hypertension, cardiogenic shock, mechanical circulatory support, TAZ gene, dilated cardiomyopathy, bridge to candidacy

The report centers on a man in his fifties who developed pump thrombosis just fourteen days after receiving a HeartMate 3 implant. The HeartMate 3, manufactured by Abbott, is a fully magnetically levitated centrifugal-flow pump designed to minimize blood trauma. Its artificial pulse feature and wide blood-flow gaps were engineered specifically to reduce the risk of thrombosis, and large randomized trials have indeed shown remarkably low rates of pump thrombosis compared with older axial-flow devices. That very success, however, means that when thrombosis does occur, clinicians may have little experience recognizing it, particularly in the early postoperative period when it is rarest and hardest to diagnose.

In this case, the patient’s initial presentation was strikingly subtle. He remained clinically stable, with no overt signs of heart failure, device malfunction, or systemic illness. Laboratory abnormalities were present but mild, lacking the dramatic elevations in plasma free hemoglobin and lactate dehydrogenase that classically accompany pump thrombosis and that form the backbone of established diagnostic algorithms. What stood out instead was the device’s own telemetry: an abrupt decline in estimated pump flow recorded by the system monitor. In a device that continuously reports flow, power, pulsatility index, and speed, a sudden fall in flow is a mechanical statement that something is obstructing the blood’s path through the pump or its conduits.

The clinical team pursued the signal aggressively. Contrast-enhanced computed tomography was performed, and the imaging raised suspicion of thrombus formation within the outflow graft, the conduit that carries blood from the pump into the aorta. Faced with a plausible mechanical obstruction and a device already showing declining performance, the surgeons proceeded to an emergent pump exchange. The decision proved prescient. Subsequent analysis of the explanted device by the manufacturer demonstrated thrombus attached to the metallic edge at the inflow aspect of the pump, the region where blood enters from the left ventricle. Notably, no thrombus was identified on the rotor or on the impeller blade surfaces, indicating that the clot had lodged at the pump’s entry rather than accumulating on its moving parts.

This anatomical detail matters because it helps explain the paradoxical laboratory picture. Hemolysis, the destruction of red blood cells that releases free hemoglobin into plasma, is most severe when thrombus interacts directly with the rapidly spinning rotor. A thrombus sitting at the inflow edge, partially obstructing inflow rather than churning against the impeller, can reduce flow while generating comparatively little shear-related blood damage. The result is a thrombotic event that evades the standard biochemical radar. The authors argue that this case exposes the limitations of relying solely on laboratory findings for early detection and prediction of early HeartMate 3 pump thrombosis, and they emphasize that an abrupt decline in pump flow, even in the absence of marked hemolysis or increased pump power, may represent an early and clinically important warning indicator.

The diagnostic challenge in continuous-flow ventricular assist devices is well documented in the literature the authors cite. Algorithms developed more than a decade ago, including the widely used approach published by Goldstein and colleagues in 2013, built diagnosis around a constellation of findings: hemolysis, rising pump power, low flow, and echocardiographic abnormalities. Later work by Scandroglio and colleagues refined the evaluation of blood flow obstructions, and a systematic analysis by Kaufmann and colleagues in 2022 catalogued thrombus formation at the inflow cannula of continuous-flow devices. Yet the HeartMate 3’s hemocompatibility profile has shifted the epidemiology. Pump thrombosis is now so uncommon that individual cases carry outsized educational value, and each one refines the community’s understanding of how the complication can present.

Early postoperative thrombosis is a particularly treacherous subset. Reported cases include thrombosis within one hour of implantation, intraoperative pump thrombosis, and acute events in the first days after surgery, as documented by Shah and colleagues, Karuppiah and colleagues, and Bunge and colleagues in separate reports. The early period is a perfect storm of prothrombotic conditions: fresh surgical surfaces, inflammatory activation, altered anticoagulation management, and a healing heart adapting to unloading by the new device. A patient in this window who develops a flow abnormality may be dismissed as having volume shifts, arrhythmia, or right heart dysfunction, all of which can alter pump flow estimates. The Kurume case demonstrates why an abrupt change should instead prompt immediate structural evaluation rather than watchful waiting.

The technology itself deserves attention in understanding why flow telemetry is such a sensitive indicator. The HeartMate 3 operates at a fixed speed set by the clinician, typically around 5,000 to 6,000 revolutions per minute, and the controller estimates flow from the relationship between power consumption and speed. When inflow becomes obstructed, the pump essentially runs against a partial vacuum, and the estimated flow falls even though the impeller continues spinning at its commanded speed. Pump power, by contrast, may change little when the obstruction is at the inflow rather than within the rotor region, which is precisely the pattern observed here. The artificial pulse, a periodic speed modulation unique to the HeartMate 3 that washes all surfaces of the pump, also influences the flow waveform, and clinicians familiar with its normal pulsatility signature can detect deviations that suggest inflow problems. In this case, the flow decline was abrupt enough to stand out against the patient’s baseline, functioning as the device’s own distress signal.

Contrast-enhanced computed tomography emerged as the decisive diagnostic step, and its role in this case reinforces a growing consensus that imaging should be part of the early evaluation of suspected device obstruction. Echocardiography remains the first-line tool for assessing inflow cannula position, ventricular size, and valve function, but it can miss thrombus within the outflow graft or at the pump’s inflow edge. Computed tomography with contrast offers a direct view of the entire blood path from ventricle to aorta, and in this patient it converted an ambiguous telemetry finding into a surgical indication. The authors’ sequence, from flow alarm to laboratory review to cross-sectional imaging to emergent exchange, effectively models the pathway they believe should be followed when abrupt flow decline appears without an obvious hemodynamic explanation.

The broader message for the mechanical circulatory support community is one of vigilance and humility about traditional markers. The HeartMate 3 has transformed outcomes for patients with advanced heart failure, with five-year results from the MOMENTUM 3 trial showing superior survival and fewer adverse events than earlier devices. But hemocompatibility is not immunity, and the rare thrombosis that does occur may present in atypical ways. The Kurume team, supported in part by a Japan Society for the Promotion of Science KAKENHI grant, closes their report with a clear clinical directive: prompt recognition and intervention are essential to prevent clinical deterioration. For the thousands of clinicians who monitor HeartMate 3 patients through daily device downloads, the case reframes a familiar number on the screen. An abrupt fall in pump flow is not a curiosity to be observed; it is a call to action, potentially the first and only warning that a life-sustaining pump is beginning to clot from the inside.

Subject of Research: Early detection of HeartMate 3 pump thrombosis using abrupt pump flow decline as an initial warning indicator

Article Title: Abrupt flow decline as the initial indicator of early HeartMate 3 pump thrombosis

Article References: Ishii, Y., Kato, T. S., Takagi, K., Sano, S., Yoshimatsu, S., Shinoda, M., Yanai, T., Shibata, T., Uehara, M., Uchimura, H., Sugihara, G., Fukumoto, Y., & Tayama, E. (2026). Abrupt flow decline as the initial indicator of early HeartMate 3 pump thrombosis. Journal of Artificial Organs, 29(4), Article 61. https://doi.org/10.1007/s10047-026-01588-1

Image Credits: AI Generated

DOI: 10.1007/s10047-026-01588-1

Keywords: HeartMate 3, pump thrombosis, LVAD, mechanical circulatory support, pump flow, hemolysis, outflow graft, computed tomography, heart failure, ventricular assist device, low flow alarm, device thrombosis