Ventricular assist devices (VADs) have revolutionized the management of advanced heart failure (HF) and significantly improved HF patients’ survival and quality of life.1 However, VAD therapy continues to be burdened by serious adverse events. Stroke has been a particularly dreaded complication as it worsens survival, dramatically decreases quality of life, and frequently is a barrier to subsequent heart transplantation.2 In recent years, newer generation VADs with improved mechanical reliability have prompted more focus on the impact of complications including stroke. How we as a field can further mitigate serious VAD complications is pivotal to the future success of long-term VAD therapy.
Neurologic events have been an important focus of recent VAD clinical trials and registries such as the Prospective, Randomized, Controlled, Un-blinded, Multi-Center Clinical Trial to Evaluate the HeartWare Ventricular Assist System for Destination Therapy of Advanced Heart Failure (ENDURANCE) Supplemental trial, the Multi-center Study of MagLev Technology in Patients Undergoing MCS Therapy With HeartMate 3 (MOMENTUM) 3 trial, and the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) registry.1,3,4 In addition, there have been a number of publications investigating various stroke risk factors in VAD patients, including suboptimal anticoagulation, device thrombosis, and infection.2,5 Yet, despite these important discoveries, our understanding of VAD-related strokes remains woefully inadequate. Stroke rates vary significantly among VAD trials, even for the same device, due to heterogeneous patient populations, stroke definitions, and patient management protocols.6 In addition, for the stroke risk factors that have been identified, their relative importance, mechanism, and interaction remain largely unknown, let alone their potential mitigation strategies.
In this issue of the ASAIO Journal, Cho et al.7 aimed to identify potential causes of acute ischemic strokes (AIS) in 477 VAD patients at a single academic center. Even though this is a single-center study with a modest number of AIS (61), the authors laudably took advantage of manageable sample size by having experts manually extract rich clinical information to gain a detailed understanding of each stroke’s cause and potential mechanisms. The authors found that all 61 AIS were associated with the perioperative period, subtherapeutic anticoagulation, suboptimal antiplatelet therapy, VAD thrombosis, or acute infection. The common thread among these stroke triggers is a prothrombotic state. This is an important observation because if most VAD strokes have identifiable and potentially preventable triggers, they can be targeted and potentially mitigated.
It is important to acknowledge the institutional practice pattern of this cohort, which included inconsistent heparin bridging to therapeutic international normalized ratio (INR), variable antiplatelet therapy, late initiation of warfarin (as long as 2 weeks), and a target INR that is significantly less than many VAD programs. In contrast, our own institutional practice, for instance, is to strictly bridge VAD patients for subtherapeutic INR (<2) with anti-Xa monitoring beginning in the early postoperative period, and to not interrupt or reverse anticoagulation for anything other than intracerebral hemorrhage or life-threatening bleeding. Meticulous surgical hemostasis at the time of VAD implantation is imperative and sets the stage to avoid a cascade of adverse events wherein surgical bleeding delays initiation of anticoagulation and then increases the risk of thrombotic complications.8 It has also been recently shown that the device display and stored data often lack granularity of important information, such as subclinical suction events that do not trigger a stored suction alarm, or identify early warning signs of device thrombosis.9
The current study by Cho et al.7 and other similar studies concerning stroke risk factors in VAD patients naturally lead to the question: how do we organize the scatter of modifiable and non-modifiable stroke risk factors into a conceptual framework that facilitates the discovery and implementation of mitigation strategies? Here we propose a comprehensive conceptual framework of stroke in VAD patients in hope of achieving this goal (Figure 1). Understanding both the modifiable and non-modifiable factors is vital for future research to reduce stroke risk in VAD patients (Figure 2).
For non-modifiable stroke risk factors, proper risk assessment and stratification are helpful for patient selection, not to mention goals of care discussions. Risk adjustment based on population baseline risk is necessary for meaningful comparison of stroke outcomes. It has been shown that each of the contemporary VAD platforms (HeartMate II, HeartMate 3, and HVAD) can achieve quite divergent stroke rates in different clinical trials based on trial design, especially patient selection.6 Thus, a VAD stroke risk model akin to the CHA2DS2-VASc model can be used to compute adjusted stroke rates in future trials and registries for a fairer comparison. In addition, while some have called for reducing stroke rate in VAD patients to “zero,” a more realistic target is perhaps a patient’s baseline stroke risk devoid of device therapy, which is elevated for many advanced HF patients with comorbidities that increase stroke risk. For example, a 65 year old male with advanced HF due to ischemic cardiomyopathy, hypertension, atrial fibrillation, and diabetes mellitus has an annual stroke risk of ~7% by the CHA2DS2-VASc risk model (equaling 0.07 event per patient-year [EPPY]).10 In comparison, the latest INTERMACS analysis showed a stroke rate of 0.18 EPPY, while the HM3 arm in MOMENTUM3 and HVAD in LATERAL had rates of 0.08 and 0.10 EPPY, respectively.6 So there is a significant proportion of VAD strokes that are simply not modifiable and yet also a significant proportion that is.
For risk factors that are modifiable such as antiplatelet therapy, anticoagulation strategy, and blood pressure control, the evidence base is expanding. For example, the ENDURANCE Supplemental trial showed that an intensive BP protocol targeting a mean arterial pressure (MAP) of less than 85 mm Hg significantly lowers stroke risk.3 It has also been shown that Doppler BP is the de facto most accurate noninvasive method to estimate MAP and should be the default BP measurement method in continuous-flow VAD patients.11 Incorporating new evidence such as this into future guidelines and device “Instruction for Use” will be helpful to drive implementation. Surgical configuration of VAD implantation is another area of active scholarly pursuit. Suboptimal VAD inflow cannula angulation or insertion depth, as well as outflow graft angulation have all been shown to increase platelet shear and residence time, potentially causing thrombosis and stroke.12–14 Our current one-size-fits-all surgical approach may soon be obsolete and perhaps replaced with a more “bespoke” VAD strategy based on careful measurement of the individual patient’s LV geometry and orientation, potentially utilizing 3D printing technology. Finally, other risk factors associated with stroke such as infections are also potentially modifiable, but we do not yet know how to best treat them in a way that improves outcomes. More research is needed on the mechanistic underpinnings of their association with stroke and how to disrupt this association.
As a field, we should obviously strive for the lowest possible VAD stroke rates; however, one can surmise that it is unrealistic to expect it to be less than the stroke rate of a “simple” HF patient with atrial fibrillation or a mechanical heart valve. Until VAD engineering truly transcends our current technologic impasse of a thrombogenic metal object surgically implanted into systemic circulation, devoid of sensors and adaptive technologies matching VAD flow to patient demand, the responsibility of avoiding suction events, high shear, and long platelet residence time, remains with the care team.
In the meantime, it is imperative that we aim to improve patient selection, continue to deliberately optimize surgical configurations, diligently measure and control BP, carefully monitor and manage anticoagulation, and avoid excess VAD speed to allow for intermittent AV opening, to ensure patients and their caregivers have the best outcomes possible from an inherently flawed therapy in its current technologic iteration.
1. Kirklin JK, Pagani FD, Kormos RL, et al. Eighth annual INTERMACS report: Special focus on framing the impact of adverse events. J Heart Lung Transplant 2017.36: 1080–1086.
2. Acharya D, Loyaga-Rendon R, Morgan CJ, et al. INTERMACS analysis of stroke during support with continuous-flow left ventricular assist devices: Risk factors and outcomes. JACC Heart Fail 2017.5: 703–711.
3. Milano CA, Rogers JG, Tatooles AJ, et al. ENDURANCE Investigators: HVAD: The ENDURANCE Supplemental Trial. JACC Heart Fail 2018.6: 792–802.
4. Mehra MR, Uriel N, Naka Y, et al. MOMENTUM 3 Investigators: A fully magnetically levitated left ventricular assist device - final report. N Engl J Med 2019.380: 1618–1627.
5. Teuteberg JJ, Slaughter MS, Rogers JG, et al. ADVANCE Trial Investigators: The HVAD left ventricular assist device: Risk factors for neurological events and risk mitigation strategies. JACC Heart Fail 2015.3: 818–828.
6. Li S, Beckman JA, Cheng R, et al. Comparison of neurologic event rates among HeartMate II, HeartMate 3, and HVAD. ASAIO J 2019. doi: 10.1097/MAT.0000000000001084 Epub ahead of print.
7. Cho SM, Hassett C, Rice CJ, Starling R, Katzan I, Uchino K. What causes LVAD-associated ischemic stroke? Surgery, pump thrombosis, antithrombotics, and infection. ASAIO J 2019.68: 775–780.
8. Angleitner P, Simon P, Kaider A, et al. Impact of bleeding revision on outcomes after left ventricular assist device implantation. Ann Thorac Surg 2019.108: 517–523.
9. Grabska J, Schlöglhofer T, Gross C, et al. Early detection of pump thrombosis in patients with left ventricular assist device. ASAIO J 2019. doi: 10.1097/MAT.0000000000001015 Epub ahead of print.
10. Lip GY, Nieuwlaat R, Pisters R, Lane DA, Crijns HJ. Refining clinical risk stratification for predicting stroke and thromboembolism in atrial fibrillation using a novel risk factor-based approach: The euro heart survey on atrial fibrillation. Chest 2010.137: 263–272.
11. Li S, Beckman JA, Welch NG, et al. Accuracy of doppler blood pressure measurement in continuous-flow left ventricular assist device patients. ESC Hear Fail2019.133: e38
12. Chivukula V, Beckman JA, Li S, et al. Left ventricle assist device inflow cannula insertion depth influences thrombosis risk. ASAIO J2019. doi: 10.1097/MAT.0000000000001068 Epub ahead of print.
13. Chivukula VK, Beckman JA, Prisco AR, et al. Left ventricular assist device inflow cannula angle and thrombosis risk. Circ Heart Fail 2018.11: e004325
Copyright © 2019 by the American Society for Artificial Internal Organs
14. Aliseda A, Chivukula VK, Mcgah P, et al. LVAD outflow graft angle and thrombosis risk. ASAIO J 2017.63: 14–23.