Anesthetic Considerations in Patients with Left Ventricular Assist Devices

With advances in healthcare, aging of the general population, and improvement in life expectancy, the prevalence of heart failure continues to rise, ensuring that all anesthesiologists can expect to care for patients with left ventricular assistive devices (LVAD). Thus, clinicians must understand the development and design of these devices as well as current guidelines regarding perioperative management.  While heart transplantation serves as the definitive treatment of refractory end stage heart failure, time to transplantation can be quite variable and often incorporates a temporizing measure such as an LVAD, in what is known as “bridge to therapy.” While waiting for a compatible organ, these patients often need noncardiac surgery to treat co-morbidities or injuries sustained in the interim. Select patients may also be treated permanently with an LVAD if they are considered poor candidates for transplantation. This approach to LVAD management is known as “destination therapy.” Studies report anywhere from 4% to one-half of all patients with a device will require noncardiac surgery while the LVAD is in place.¹

Prior to 2010, a significant proportion of the devices in use were “pulsatile-flow” type devices, however since that time, the majority of devices distributed have been designed for “continuous flow.” Newer designs are incorporating methods to recreate pulsatility in the peripheral vasculature that is lost with LVAD implantation. The current iterations in production are the HeartMate II and HeartWare (HVAD). The HVAD is not currently approved for destination therapy, thus the majority of devices in use are HeartMate II. Both devices operate using a magnetic field which rotates an impeller to drive blood flow through a vascular graft to the aorta. While there are significant differences in their design, both devices are common in that they are: implanted in the apex of the LV, assist cardiac output via continuous flow through ascending aorta, powered via a driveline that connects percutaneously to AC or DC power supply and the VAD controller. Flow is set at the controller and is commonly displayed in revolutions per minute (RPM). In real-time, flow through the device is dependent on preload, afterload, and actual RPM (which may differ from set RPM at the controller based on acute changes in physiology). LVADs are preload-dependent, meaning any increase in preload will be reflected as increased flow. Physiologic changes such as hypovolemia, RV failure, cardiac tamponade, or pulmonary hypertension will significantly impact preload, limiting flow through the VAD. If preload falls too low, the device may actually cause collapse of the ventricle, leading to arrythmia or hemodynamic compromise by creating a “suction” effect on the left ventricle. Changes in afterload also affect the LVAD similarly to the native heart. Decreased afterload will allow for increased flow, however in the setting of heart failure, end organs become ischemic if mean arterial pressure remains low. Therefore, a MAP between 70-90 mmHg provides optimal balance between forward flow through the VAD and end-organ perfusion. Flow is also dependent on the speed of the VAD; however, increasing the speed can also lead to emptying of the LV and overload of the aortic root, which may complicate native pulsatile flow across the aortic valve. In such a situation, all flow preferentially proceeds through the VAD into the aorta via the graft, while the aortic valve remains closed throughout the cardiac cycle. This can overload the systemic side of the valve, and has been associated with the development of aortic insufficiency, valve thrombosis, and the formation of arteriovenous malformations which have been known to cause gastrointestinal bleeding in this patient population – especially in the setting of chronic anticoagulation required for proper functioning of the device.

Given the high risk of pump thrombosis while the LVAD is in place, anticoagulation is essential. However, regimens may differ between institutions. Patients are commonly maintained on warfarin and an anti-platelet agent and bridged to heparin perioperatively – albeit, this is case and situation dependent. The reported incidence of thrombosis requiring pump exchange is around 2-12%. Modifiable risk factors for pump thrombosis are active infection and pump speed less than 8600 RPM.⁸ In the setting of a continuous flow VAD, bleeding has been reported more frequently than thrombosis, again with GI bleeding accounting for more than half of these events. Given the association of arteriovenous malformation and GI bleeding seen commonly in VAD patients, they are frequently anesthetized in the endoscopy suite to evaluate their GI lesions. LVAD patients also appear to develop an acquired type 2A Von Willebrand’s disease related to the high shear stress the VAD generates at the cellular/molecular level, resulting in loss of the high molecular weight multimers (HWMWs) responsible for binding and hemostasis. While this coagulopathy is seen in all VAD patients and resolves quickly following device removal, not all patients will have hemorrhagic complications.

Regarding the perioperative management of patients with an LVAD, there are several important considerations at each phase. Location is of concern, as a center with a heart transplant program is likely to have staff who are familiar with LVADs. Additionally, they will have a cardiothoracic surgeon and perfusionist on call in case of an emergency or complication with the device. Staffing of cases by non-cardiac anesthesiologists occurs more than 40% of the time³, again emphasizing a need for LVAD training and continuous education for all anesthesia staff. In major cases with high risk patients, or those maintained on vasoactive infusions, it is ideal for a cardiac anesthesiologist to be assigned to the case. The most commonly scheduled procedures for patients with an LVAD are tracheostomies, upper/lower endoscopies, and vascular access cases; however, it is not uncommon for these patients to present for major noncardiac surgeries.⁹ Procedural room size and design also deserves consideration, as the LVAD controller and drivelines must have ample space and cannot be blocked or under tension. Ideally, the pump will be connected to AC power for the duration of the procedure, in order to preserve battery life in case of intraoperative power failure. Batteries should remain readily available to the anesthesia provider.

Preoperatively, the patient’s anticoagulation regimen should be verified and discussed with the surgeon and cardiologist if possible. Risk of perioperative hemorrhage should be balanced with that of pump thrombosis and any other comorbidities requiring anticoagulation – such as atrial fibrillation or venous thromboembolism. No clear standard exists regarding perioperative anticoagulation in these cases; that said, according to the International Society of Heart and Lung Transplantation, warfarin and anti-platelet therapy should be continued if the risk of procedural bleeding is low in elective cases. If warfarin must be held, heparin bridging should be considered as it has been shown to decrease the incidence of pump thrombosis.⁸ In some institutions, therapy is held without bridging, especially in high risk cases such as neurosurgery or ophthalmic procedures. Yet, the risks of this must be discussed thoroughly with the multidisciplinary team caring for the patient. For emergency procedures, reversal of anticoagulation may be considered, but again must take into account the type/nature of surgery, patient comorbidities, risk factors for procedural hemorrhage versus pump thrombosis, and risk of thrombosis at other sites.

The patient’s right ventricular (RV) function should also be assessed preoperatively in the case of elective surgery. Pertinent echocardiographic studies as well as a thorough history and chart review are necessary at minimum to assess for any prior inotropic support or intervention to treat RV failure. Pre-existing RV dysfunction heralds an increased risk of intraoperative RV complications. In general, RV dilation is a sign that the cardiac circuit is more susceptible to volume overload and under-resuscitation, both of which may lead to intraoperative pump dysfunction and hypoperfusion. Preoperative RV assessment is essential in the setting of planned high-risk surgery where blood loss is expected to be significant, as it may guide the anesthesiologist in planning for invasive monitoring and/or central access. Balancing resuscitative efforts for optimal perfusion may require intraoperative transesophageal echocardiography or central pressure and cardiac output monitoring to differentiate between hypovolemia and RV dysfunction as possible causes of low preload.

Managing the patient with an LVAD intraoperatively requires an understanding of the changes in cardiac physiology seen in advanced heart failure, physiology of the LVAD circuit, and VAD monitoring parameters, as well as their clinical significance. Noninvasive blood pressures and pulse oximetry may be difficult to obtain depending on the degree of pulsatility present in the peripheral vasculature. One study reports success with a blood pressure cuff in around 50% of cases, with only a mean arterial pressure being displayed in 40% of those cases.⁴ The narrow pulse pressure physiology makes invasive monitoring with an arterial catheter useful if not necessary in many cases but also makes placement more challenging. Cerebral oximetry is also useful in select cases as a surrogate marker of cardiac output. Considerations to bear in mind when deciding whether to use invasive monitoring include the likelihood of needing multiple future surgeries, intraoperative access to extremities, and difficulty of cannulation intraoperatively in the setting of low pulsatility. If the case is relatively low risk and noninvasive monitoring is possible, it may be prudent to forego placing arterial lines as this may complicate future attempts for more high-risk surgeries.

Monitoring during procedures requires an understanding of the LVAD monitor and the impact various anesthetic interventions will have on device parameters. Power, RPM, pulsatility index, and flow are displayed on the LVAD monitor; however, flow is actually a calculated parameter and cannot be used as a direct surrogate for cardiac output for a number of reasons. The power used by the device is generally proportional to RPM and flow. Yet, there are situations such as pump thrombosis where increasing power may be a sign of poor flow. Pulsatility or the pulsatility index, is trended over the course of most major cases, as loss of pulsatility or an index less than 3 can be a sign of poor pump function. Pulsatility index (PI) is calculated as [10x(Max Flow – Min Flow)/Average Flow] and normally ranges between 3 and 6. Drops in preload and/or afterload are often reflected on the monitor as loss of pulsatility, especially when RPM is stable. Of note, PI will decrease in the setting of increasing power or RPM because the difference in maximum and minimum flow decreases. In general, lower PI is consistent with increased support from the VAD, and in the absence of changes to RPM also correlates with hypotensive episodes.

Intraoperative goals for the LVAD patient are aimed at maintaining preload and afterload and preventing right ventricular strain. Mean arterial pressure should be kept between 70-90 mmHg, as previously stated, for optimal hemodynamics , end-organ perfusion, and VAD performance. Baseline pulsatility should be noted and thereafter PI should be maintained >3. Any drops in pulsatility should be correlated with medical/surgical events and addressed quickly, with the understanding that precipitous drops will also lead to noninvasive monitor malfunction. Measures should be taken to avoid increases in pulmonary vascular resistance which will affect preload and pulsatility. Euvolemia, normoxia, normocarbia, normothermia, adequate analgesia and anxiolysis, and spontaneous ventilation are all effective in controlling RV strain. When positive pressure ventilation is required, lung protective tidal volumes and low PEEP should be used. In patients with pre-existing RV dysfunction, transesophageal echocardiography is useful intraoperatively to evaluate changes in pulsatility that accompany increased CVP. However, de novo RV dysfunction is rarely reported intraoperatively.⁹ At any rate, hemodynamic support with inotropes, vasopressors, or pulmonary vasodilators should be readily available should hemodynamic support be required intraoperatively, especially in the rare event of pump failure.

Regarding potential complications, failure of the device has been reported in 0.1% of cases, however the anesthesiologist should always be mentally prepared and equipped for this possibility. Of note, a clinically significant aortic insufficiency typically accompanies pump failure due to the altered aortic anatomy present in LVAD. Without a functioning VAD, cardiac output through the aortic valve backflows down the graft into the apex of the left ventricle and may be as high as 1-2 L/min.  Arrythmias, when sustained, are be poorly tolerated by the patient with an LVAD. Attention should be turned to electrolytes as well as the possibility of a “suctioning” event precipitated by low preload or afterload.⁹ Finally, should advanced cardiovascular life support be required, there is concern that chest compressions may damage or dislodge the LVAD; however, the American Heart Association currently recommends proceeding with chest compressions for the VAD patient in circulatory failure.¹¹

In summary, end stage heart failure and left ventricular assist devices have become more prevalent, requiring all anesthesiologists to become familiar with the perioperative management of these patients. Thorough preoperative assessment and planning is essential, as these patients live within a fairly narrow physiologic window. Risk assessment and perioperative anticoagulation planning is always necessary to balance the goals of pump thrombosis prevention and surgical hemostasis. In addition, monitoring these patients is often challenging due to the lack of pulsatile blood flow. However, meticulous preload and afterload management are essential to prevent decompensation. While there are several important considerations for the anesthesiologist, reported complications are rare and LVAD design continues to be improved upon, ensuring anesthesia providers will be challenged with caring for patients with an ventricular assist device.

References

1. Ahmed M, Le H, Aranda JM, Klodell CT. Elective noncardiac surgery in patients with left ventricular assist devices. J Card Surg. 2012;27(5):639-642. Accessed Jan 8, 2020. doi: 10.1111/j.1540-8191.2012.01515.x.
2. Arnaoutakis GJ, Bittle GJ, Allen JG, et al. General and acute care surgical procedures in patients with left ventricular assist devices. World J Surg. 2014;38(4):765-773. Accessed Dec 16, 2019. doi: 10.1007/s00268-013-2403-0.
3. Barbara DW, Wetzel DR, Pulido JN, et al. The perioperative management of patients with left ventricular assist devices undergoing noncardiac surgery. Mayo Clin Proc. 2013;88(7):674-682. Accessed Jan 8, 2020. doi: 10.1016/j.mayocp.2013.03.019.
4. Bennett MK, Roberts CA, Dordunoo D, Shah A, Russell SD. Ideal methodology to assess systemic blood pressure in patients with continuous-flow left ventricular assist devices. J Heart Lung Transplant. 2010;29(5):593-594. Accessed Jan 10, 2020. doi: 10.1016/j.healun.2009.11.604.
5. Chung M. Perioperative management of the patient with a left ventricular assist device for noncardiac surgery. Anesthesia & Analgesia. 2018;126(6):1839. https://journals.lww.com/anesthesia-analgesia/fulltext/2018/06000/Perioperative_Management_of_the_Patient_With_a.14.aspx. Accessed Jan 8, 2020. doi: 10.1213/ANE.0000000000002669.
6. Dalia AA, Cronin B, Stone ME, et al. Anesthetic management of patients with continuous-flow left ventricular assist devices undergoing noncardiac surgery: An update for anesthesiologists. Journal of Cardiothoracic and Vascular Anesthesia. 2018;32(2):1001-1012. https://www.sciencedirect.com/science/article/pii/S1053077017309205. doi: 10.1053/j.jvca.2017.11.038.
7. Kollmar JP, Colquhoun DA, Huffmyer JL. Anesthetic challenges for posterior spine surgery in a patient with left ventricular assist device: A case report. A&A Practice. 2017;9(3):77. https://journals.lww.com/aacr/fulltext/2017/08010/Anesthetic_Challenges_for_Posterior_Spine_Surgery.4.aspx. Accessed Dec 14, 2019. doi: 10.1213/XAA.0000000000000531.
8. Maltais S, Kilic A, Nathan S, et al. PREVENtion of HeartMate II pump thrombosis through clinical management: The PREVENT multi-center study. J Heart Lung Transplant. 2017;36(1):1-12. Accessed Jan 10, 2020. doi: 10.1016/j.healun.2016.10.001.
9. Mathis MR, Sathishkumar S, Kheterpal S, et al. Complications, risk factors, and staffing patterns for noncardiac surgery in patients with left ventricular assist devices. Anesthesiology. 2017;126(3):450-460. Accessed Jan 10, 2020. doi: 10.1097/ALN.0000000000001488.
10. Nelson EW, DO, Heinke T, MD, Finley A, MD, et al. Management of LVAD patients for noncardiac surgery: A single-institution study. Journal of Cardiothoracic and Vascular Anesthesia. 2015;29(4):898-900. https://www.clinicalkey.es/playcontent/1-s2.0-S1053077015000506. doi: 10.1053/j.jvca.2015.01.027.
11. Peberdy MA, Gluck JA, Ornato JP, et al. Cardiopulmonary resuscitation in adults and children with mechanical circulatory support: A scientific statement from the american heart association. Circulation. 2017;135(24):e1115-e1134. Accessed Jan 10, 2020. doi: 10.1161/CIR.0000000000000504.
12. Uriel N, Han J, Morrison KA, et al. Device thrombosis in HeartMate II continuous-flow left ventricular assist devices: A multifactorial phenomenon. J Heart Lung Transplant. 2014;33(1):51-59. Accessed Jan 8, 2020. doi: 10.1016/j.healun.2013.10.005.