A rare and potentially fatal complication of long-term mechanical heart support has drawn attention to an overlooked source of clinical evidence: the digital log produced by an implanted artificial heart pump. In a case reported in the Journal of Artificial Organs, physicians described a late-onset obstruction of the outflow graft in a patient supported by an EVAHEART2 left ventricular assist device. The obstruction was associated with infection by Candida albicans, an organism capable of producing invasive, persistent disease in patients with implanted cardiac hardware. The report’s central message is that subtle changes in pump behavior may reveal a developing obstruction before conventional alarms or obvious clinical deterioration appear.
Left ventricular assist devices, or LVADs, are used to support patients with severe heart failure when the heart can no longer maintain adequate circulation. In the EVAHEART2 system, blood is drawn from the left ventricle through an inflow cannula and propelled into the aorta through an artificial outflow graft. The pump therefore creates a mechanical pathway that must remain unobstructed for blood to reach the body. Any narrowing or blockage in this route increases resistance to flow and can reduce effective circulatory support. Because the device is continuously connected to the patient’s bloodstream and vascular system, an obstruction may quickly become a life-threatening emergency, potentially leading to low cardiac output, organ dysfunction, thrombosis, or sudden circulatory collapse.
Outflow graft obstruction is considered uncommon, but its causes are diverse. The graft can become compressed, twisted, kinked, or narrowed by external tissue. Blood clots may accumulate within the conduit, while inflammatory tissue or infection can gradually reduce the internal diameter. Fungal infection is especially concerning because it may develop insidiously and can be difficult to diagnose using routine tests. Candida albicans can adhere to artificial surfaces and form biofilms, structured communities of microorganisms protected by a matrix that reduces their vulnerability to immune defenses and antimicrobial treatment. When such an infection involves an implanted circulatory-support system, the resulting mass or inflammatory reaction may progressively interfere with blood passage through the graft.
The EVAHEART2 presents a particular monitoring challenge because it does not provide a conventional low-flow alarm. Many LVAD systems alert clinicians when estimated pump flow falls below a predefined threshold. Such alarms can serve as an immediate warning that the pump is not receiving or delivering enough blood. Without that feature, clinicians caring for EVAHEART2 patients must integrate symptoms, imaging, laboratory findings, pump parameters, and trends in device performance. A single measurement may remain within an apparently acceptable range even while a mechanical problem is developing. The Japanese team’s report emphasizes that the direction and speed of change in the device’s recorded data may be more informative than an isolated value.
The investigators identified a characteristic pattern in the EVAHEART2 log data. Over time, the maximum pump speeds recorded during operation gradually declined, accompanied by a reduction in maximum power consumption. At the same time, the minimum pump speeds and minimum power values gradually increased. In simple terms, the distance between the pump’s upper and lower operating states became progressively narrower. This convergence suggested that the device was losing the ability to vary its performance normally in response to changing circulatory conditions. The pattern did not necessarily announce an abrupt failure; instead, it reflected a slow erosion of the pump’s operating range, consistent with increasing resistance somewhere in the outflow pathway.
Understanding why this pattern matters requires a look at how a centrifugal blood pump responds to resistance. The pump’s rotating impeller transfers energy to the blood, generating pressure that drives flow through the outflow graft and into the aorta. If the graft begins to narrow, the pump must work against a higher pressure load. Depending on the control strategy and the patient’s condition, the pump may alter its speed to maintain a target level of support or physiological pulsatility. As obstruction worsens, the system may be unable to reach its previous maximum speed without exceeding operational limits, while its lower operating state may also shift upward. Power consumption reflects the electrical energy required to rotate the pump and can change with both speed and hydraulic load. Trends in speed and power can therefore act as indirect indicators of altered blood-flow resistance.
The case is important because the log-data signal appeared to offer an opportunity for earlier recognition of a problem that might otherwise remain hidden. Outflow graft obstruction can be difficult to distinguish from other causes of worsening heart failure, including changes in volume status, right-heart function, arrhythmia, infection, or disease progression. Imaging studies such as computed tomography, echocardiography, angiography, or specialized pump testing may be required to localize the obstruction and determine its severity. However, clinicians must first suspect that a mechanical complication exists. A gradual, reproducible change in the relationship between pump speed and power could provide that initial trigger, prompting closer investigation before the patient develops severe symptoms or an abrupt reduction in support.
The fungal origin of the obstruction adds another layer of urgency. In patients with durable LVADs, infection is already a major concern because driveline sites, cannulas, grafts, and pump components can provide surfaces for microbial attachment. Fungal infections are less common than bacterial infections but are often associated with difficult treatment, prolonged illness, and a high risk of recurrence. A Candida infection may not present with a dramatic fever or a clear bloodstream infection at the beginning. Instead, the organism may persist around prosthetic material while the mechanical consequences emerge gradually. In this case, the obstruction demonstrates that infection can affect not only the biological safety of an LVAD but also its hydraulic performance, creating a problem that combines infectious disease, cardiovascular surgery, and biomedical engineering.
The authors’ observation also illustrates how artificial-heart technology produces a new form of clinical data. Modern support systems continuously record variables that are not directly visible at the bedside, including rotational speed, electrical power, estimated flow, alarm states, and changes in control behavior. These logs can be viewed as a kind of mechanical vital-sign record. Their value is greatest when analyzed longitudinally, because slowly evolving abnormalities may be missed when data are reviewed only after a crisis. For EVAHEART2, the absence of a low-flow alarm makes systematic log review particularly relevant. Regular assessment of maximum and minimum speed and power trends could help care teams recognize deviations from an individual patient’s established baseline and investigate them before the obstruction becomes critical.
The report does not suggest that every change in pump speed or power proves an outflow graft obstruction, nor that log analysis can replace clinical examination and imaging. Pump parameters are influenced by preload, afterload, blood pressure, volume status, right-ventricular performance, device settings, and other physiological factors. Nevertheless, the case proposes a practical warning signature: a gradual fall in the device’s maximum speed and power together with a gradual rise in its minimum speed and power. By bringing this pattern to wider attention, the investigators argue that device-log analysis should become part of routine surveillance for EVAHEART2 patients. The broader lesson extends beyond one pump model: in an era of increasingly sophisticated mechanical circulatory support, the earliest warning of danger may be hidden not in a dramatic alarm, but in a slowly changing digital trace.
Subject of Research: Outflow graft obstruction caused by Candida albicans after EVAHEART2 left ventricular assist device implantation, and the use of device-log analysis for early detection.
Article Title: A case of outflow graft obstruction after EVAHEART2 implantation: role of device log analysis
Article References: Muratsuji Y, Kawamura A, Yoshioka D, et al. “A case of outflow graft obstruction after EVAHEART2 implantation: role of device log analysis.” Journal of Artificial Organs, volume 29, article 32, 2026. Related references include Hoermandinger C et al., Artificial Organs (2023); Peters CJ et al., Journal of Clinical Medicine (2023); Agrawal A et al., ESC Heart Failure (2022); Grüger T et al., Journal of Thoracic and Cardiovascular Surgery (2018); and Yamazaki K et al., General Thoracic and Cardiovascular Surgery (2007).
Image Credits: AI Generated
DOI: 10.1007/s10047-026-01558-7
Keywords: EVAHEART2, outflow graft obstruction, Candida albicans, left ventricular assist device, mechanical circulatory support, device log analysis, fungal infection, centrifugal blood pump.
Tags: artificial heart device failure indicatorsCandida albicans infection in cardiac devicesdigital logs in artificial heart monitoringearly detection of mechanical heart pump issuesEVAHEART2 mechanical heart support complicationimpact of infections on LVAD functionimportance of subtle pump behavior changesinfection-related outflow graft obstructionlate-onset pump obstruction detectionlong-term LVAD complication managementLVAD outflow graft obstructionsigns of outflow graft narrowing
