Postoperative VAD Management: Operating Room to Discharge and Beyond: Surgical and Medical Considerations





Perioperative management


This chapter focuses on the early postoperative care of patients with ventricular assist devices (VADs). Appropriate patient selection and optimal implantation techniques greatly simplify postoperative management. Indeed, some problems created by suboptimal implantation techniques are difficult or impossible to correct, short of a return to the operating room. This chapter commences with a discussion of operative factors that influence the early intensive care unit (ICU) course of a VAD patient, reviews early postoperative challenges in the ICU and their management, and addresses the further hospital course leading up to discharge and considerations of postdischarge management once the patient first returns to the community. We start in the operating theatre.




Considerations in the operating room relevant to subsequent ICU care


Appropriate de-airing of the pump and heart is crucial to a successful outcome. The surgeon carries out the de-airing maneuvers in conjunction with the perfusionist and the anesthesiologist while monitoring the transesophageal echocardiogram. A substantial amount of air can be removed from the patient’s left ventricle (LV) and VAD before completing the final connections of the system (i.e., 1—outflow graft anastomosis to aorta or 2—outflow graft connection to the pump housing). If the final step is to anastomose the outflow graft to the aorta, attaching a sump sucker to the outflow graft and then filling the beating heart with blood pushes blood from the heart through the VAD and into the sump sucker. If the final step is to attach the outflow graft to the pump housing, a vent tube can be attached to the pump outlet thread protector. The surgeon and perfusionist fill the beating heart with blood to push blood and air into the vent line.


While the pump and heart are in this initial phase of de-airing, the lungs should be gently ventilated with the patient in a head-down (Trendelenburg) position with a second venting device aspirating blood from the aortic root. While this initial de-airing occurs, the surgical team can accomplish other tasks, including placement of the percutaneous driveline and positioning of the outflow graft in the desired position (e.g., along the diaphragmatic surface of the pericardium and lateral to the right atrium), with a few sutures to prevent migration.


The final de-airing takes place after the pump connections are complete. It is best for each team to develop their own protocol for de-airing; however, the general plan contains the following features. The process begins by fully reinflating the lungs and setting the ventilator to resume regular respirations. The rhythm of the heart is optimized using pacing and medications as necessary. Ideally, the patient will be in a normal sinus or paced atrial rhythm. The perfusionist then decreases venous return and begins partial cardiopulmonary bypass. When the surgeon feels that there is adequate blood in the LV, the pump is powered up and rotor rotation is started, but with a cross-clamp in place on the outflow graft. Blood and air are aspirated from the outflow graft proximal to the clamp via a venting needle. After completing the initial de-airing, the outflow graft is slowly unclamped while monitoring the LV and ascending aorta for air. The vent in the ascending aorta captures any air that escapes the outflow graft vent. Shaking the heart and forcefully ventilating the lungs dislodge smaller bubbles of air trapped along the walls of the heart and pump. After completing the de-airing sequence with the patient in a head-down position, the maneuvers are repeated with the patient in a supine (flat) position.


Air that enters the cerebral circulation may cause brain injury ranging from confusion to more severe focal and global brain injury (e.g., multiple small regions of stroke). Air entering the ascending aorta and right coronary artery (RCA) may cause—at least transiently—right ventricle (RV) dysfunction, resulting in immediate RV dilation and free wall hypokinesis. At this point, it is prudent to go back on full cardiopulmonary bypass in order to push the air through the right coronary circulation and recover RV function.


As the echocardiographer and surgeon monitor for air, they can also assess native cardiac function and the effects of the VAD on the heart. This includes determining valve function (e.g., assessments for mitral, aortic, and tricuspid valve insufficiency) and the position of the ventricular septum relative to the LV and RV cavities. Another important item to check with echocardiography is the position of the inflow cannula relative to the LV septum and free wall. Ideally, the orifice of the inflow cannula will point posteriorly, directly at the mitral valve orifice. Doppler interrogation of the LV cavity confirms nonturbulent flow directly from the mitral valve into the VAD. The inflow cannula should not be angled toward the intraventricular septum, but rather parallel to the septum to avoid septal suction, turbulent flow, and potential ventricular tachycardia.


As the LV fills and begins to eject blood through the aortic valve, the surgeon increases pump rotor speed (revolutions per minute) to a point where the VAD provides adequate systemic blood flow without pulling the ventricular septum toward the LV. Upon achieving this balance of native ventricular function and pump function, the flow of the cardiopulmonary bypass circuit is decreased and eventually discontinued. At this point, the patient’s cardiac output is a combination of the native LV ejection and left VAD (LVAD) pump flow. If systemic flow is adequate and the other echocardiographic indicators of VAD function are satisfactory (e.g., inflow cannula position, septal position, and absence of air in the heart or aorta), the surgeon removes the cardiopulmonary bypass cannulas as the anesthesiologist administers protamine.


Pharmacological management of circulation early after VAD implantation focuses primarily on RV function. This includes inotropic medications (often milrinone, which simultaneously reduces RV afterload) to stimulate RV contractility, vasopressors to maintain adequate systemic blood pressure required for right coronary artery perfusion, and inhaled pulmonary vasodilators (often inhaled prostanoids as a less expensive alternative to inhaled nitric oxide) to reduce pulmonary vascular resistance without causing V/Q mismatch and resultant hypoxemia or systemic hypotension. If blood pressure is low or the patient has poor renal function, low-dose epinephrine or dobutamine may be used for inotropic support of the RV in place of milrinone. Some centers use both low-dose epinephrine and milrinone for support.


After administering protamine, the surgical team typically turns its attention to obtaining hemostasis. The surgeon must also decide whether to close the chest primarily or leave the sternum unclosed and either packed or treated with a vacuum-assisted wound management device. When the wounds are closed or sealed, the implantation team transports the patient to the ICU. If there is significant RV dysfunction, either due to a marginal RV prior to surgery or an injury to the RV intraoperatively, the surgeon may leave the chest open to assist RV function or implant a temporary right VAD (RVAD) either surgically or percutaneously for initial RV support.




Early postimplantation ICU care


The early postimplantation care of patients with a VAD is similar in many ways to the care of other postcardiac surgery patients who arrive from the operating room. The following section focuses on the unique aspects of caring for VAD patients.


LVAD placement among patients with end-stage heart disease has a high prevalence of postoperative bleeding, although the incidence of bleeding in the first 12 months postimplantation has significantly decreased from the 2008 to 2010 era (when compared to the 2011–2013 era). The causes of postimplantation bleeding include scarring and adhesions from prior cardiac operations, as well as chronic congestion of the liver that may result in liver dysfunction and resultant coagulopathy. Known cirrhosis would typically be a contraindication to durable LVAD implantation. It is difficult to measure hepatic reserve in a patient with chronic congestion from heart failure. Low serum protein and albumin levels, along with an elevated international normalized ratio (INR), suggest that hepatic synthesis of clotting factors in response to incisional bleeding and cardiopulmonary bypass will be diminished and that a fibrinolytic state may be prolonged after surgery. Preoperative right upper quadrant ultrasound can be suggestive of cirrhosis, although liver biopsy is considered the gold standard. Some groups routinely use thromboelastography to monitor the rate of clot formation, clot strength, and fibrinolytic activity. However, many groups rely on prothrombin time/partial thromboplastin time (PTT)/INR and platelet count to guide blood product and blood factor therapy early after VAD implantation.


Another useful maneuver for managing postimplantation coagulopathic bleeding is packing of the mediastinum to encourage clotting and allow time for the coagulopathy to resolve. This can be accomplished by placing sponges or laparotomy pads around cannulation sites, anastomotic sites, and against cut tissue edges followed by closure of the skin. Another useful maneuver is to pack the mediastinum as described earlier and then apply a vacuum-assisted wound closure device (e.g., Wound V.A.C.; KCI, San Antonio, TX). Use of delayed sternal closure minimizes the chance of tamponade by leaving the sternal halves separated and optimizing drainage of the mediastinum. Of note, the vacuum module for negative pressure wound management systems can be set to a variety of levels measured in mm Hg (typically ranging from − 50 to − 125 mm Hg) that exceed the maximum suction that a chest tube drainage system can provide (typical maximum suction, − 40 cm water). To ensure that one system does not prevent drainage from the other, it is important to balance the vacuum-assisted wound closure and chest tube drainage systems at the bedside. The conversion factor is 1 cm H 2 O = 0.736 mm Hg. Thus, the maximal suction for a typical chest drainage system (e.g., Pleur-evac; Teleflex, Morrisville, NC) is 40 cm H 2 O, which equals 29.4 mm Hg. This is below the preset suction options for vacuum modules. A reasonable starting point is 75 mm Hg vacuum for the mediastinal sponge dressing. This provides 102 cm H 2 O suction at the top of the wound sponge, but less in the deeper planes of the mediastinum. Typically, 75 mm Hg of vacuum still allows some of the chest drainage to exit via the chest tube drainage system. Low negative pressures (50–100 mm Hg suction) provide chest wall stability that is similar to stability at higher vacuum settings, so higher vacuum settings do not provide an advantage in this regard. Of note, high vacuum settings can cause air leaks at the edge of the dressing or organ injury. Complications from prolonged use of negative pressure wound therapy in the setting of poststernotomy mediastinitis are uncommon (7% in one series ). Complications from short-term (1–3 days) use of negative pressure wound therapy include clogging of the drainage tube with clot and loss of the air-seal at the wound edges. In our experience, these problems are rare and easily remedied.


Use of packing sponges or a vacuum system requires a return trip to the operating room for removal of the dressings and packing sponges. Mediastinal reexploration provides an opportunity for removal of mediastinal thrombus. Clot removal reduces the chances of late infection, since bacteria, especially Staphylococcus , have an affinity for growing in blood clots. This is due to the limited number of viable phagocytes in clots and the presence of nutrients, including free hemoglobin, that support bacterial growth. Use of an irrigation system (e.g., Pulsavac; Zimmer, Inc., Warsaw, IN) is helpful is removing residual thrombus from the surfaces of the mediastinal tissues and from the external surfaces of the pump. The pulsing irrigation helps wash out fibrin, necrotic debris, and any bacteria that may have inoculated the wound. A small amount of Betadine soap (1–2 mL per liter saline irrigation) (Betadine; Purdue Pharma, LP, Stamford, CN) helps remove any bacteria that adhere to tissue surfaces and disrupts small colonies of bacteria that are in the early stages of biofilm development.


RV failure that limits LVAD performance is a relatively common and serious problem. Placement of an RVAD before the sequelae of low cardiac output development will improve patient survival. Typically, the RV performance improves as pulmonary vascular resistance diminishes and the heart recovers from acute surgery. Early after LVAD implantation, pharmacologic measures to decrease pulmonary vascular resistance are helpful. If these do not produce the desired result within a few hours, placement of an extracorporeal centrifugal RVAD such as the Centrimag centrifugal flow extracorporeal VAD (Abbott Inc., Abbott Park, IL) or a novel percutaneously inserted RVAD should be considered before the patient is exposed to prolonged marginal circulatory support, as well as high doses of inotropes and vasopressors. One such novel percutaneous device is the Impella RP (Abiomed, Danvers, MA), shown to be feasible in a very small prospective study that included a post-LVAD, RV failure cohort. Another such device is the TandemLife pump with Protek Duo cannula (TandemLife, Pittsburg, PA) that is inserted as a single dual-lumen cannula through the internal jugular vein, as opposed to the Impella RP, which is inserted through the femoral vein, thus offering greater mobility. Some centers routinely bring their LVAD patients out of the operating room on inhaled prostanoids (or inhaled nitric oxide) and low-dose inotropes for empiric RV support and wean the support as tolerated by the individual patient. The efficacy of this strategy has not been proven. Associations between the length of inotropic support of the RV after LVAD implantation and death awaiting transplant have been demonstrated in retrospective analyses ; however, this may represent a population with severely compromised RV function prior to LVAD implantation.


Surgical techniques for implanting a second durable LVAD in the RV position have been described. Placement of a second durable device within the chest can be challenging, since the volume of the two pumps within the chest may compromise cardiac and pulmonary function. The combined volume of two implanted devices can make sternal closure challenging, unless the patient’s chest is relatively large. The flow of the two pumps must be balanced to avoid high mean pressures in the left or right atrium due to overcirculation of one pump relative to the other. Use of right and left atrial pressure monitoring catheters aid in appropriately setting the speed of the pumps in the early postoperative period. The ability of the native ventricles to balance the right and left circulation and the load-responsiveness of rotary pumps is helpful in achieving a right-left balance. Placement of the RVAD inflow cannula is difficult if the patient has a small RV or a fragile right atrium. Therefore, placement of the inflow cannula is best individualized to each patient. It is important to remember that all durable centrifugal pumps were designed for use in the systemic circulation. As such, the optimal pump speed for durable LVADs in the right heart position is unknown. Some groups use constricting sutures on the RVAD outflow graft to mimic systemic vascular resistance. The benefit of outflow graft constricting sutures has not been proven, although outflow graft constriction may improve the long-term performance of centrifugal pumps in the pulmonary circulation by maintaining optimal rotor position in the stator housing. Importantly, in the short-term, constricting the outflow graft adds resistance such that running the pump within its optimal speed range will not result in excess cardiac output overloading the LV or over-perfusing the pulmonary circulation. The most extensive experience in using two durable LVADs for a biventricular assist device configuration comes from the implantation of dual HeartWare HVAD (Medtronic, Minneapolis, MN) devices. Over 600 such cases have been reported in the Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS). Recently, a case series of similar implantation of two pumps using the HeartMate 3 centrifugal flow pump has been published consisting of a series of 18 total patients. As opposed to positioning the HVAD on the right side, where the outflow graft must often be constricted to simulate systemic vascular resistance and allow the pump to run at speeds (rpm > 2400) within its required parameters to avoid touch-down events, this does not seem to be required of the HeartMate 3 device ( Fig. 12.1 ). Because it employs full magnetic levitation (MagLev), the HeartMate 3 rotor will not have touch-down events at any speed. Regardless, patients requiring biventricular durable LVADs have a significantly increased risk of early death or death prior to transplantation.




Fig. 12.1


Picture of the HeartMate 3 left ventricular assist system pump.


The role of tricuspid valve repair at the time of LVAD placement remains controversial because improved RV contractility, lower pulmonary artery pressure, and falling pulmonary vascular resistance over time may decrease tricuspid regurgitation (TR) following LVAD implantation even if tricuspid repair is not performed at the time of implantation. During the 2000–2013 timeframe, most, but not all, studies suggested a benefit for tricuspid valve repair with an undersized annuloplasty ring among those with moderate to severe TR, and this was the common practice. More recent data suggest that even severe TR will improve over time as the patient’s pulmonary pressures come down and the heart remodels. The extra time on bypass to repair the tricuspid valve may be worth the risk in those at high risk for development of right heart failure (RHF), but not necessarily in all comers. Some centers are moving away from this practice in patients with mild and moderate TR, performing tricuspid valve repair or replacement only among those with tenuous RV function and severe TR. The quality of the data, both for and against tricuspid repair, is limited by the single-center design in most cases, although there were a few studies using INTERMACS data. Nonetheless, the studies’ retrospective design and the possibility of selection bias potentially confound the study results. Whether the procedure is beneficial among LVAD patients with moderate to severe TR can only truly be answered by a prospective, randomized trial.


Measures during the perioperative period to prevent infection include prophylactic antibiotics and appropriate driveline dressing with aseptic technique. The 2017 International Society for Heart and Lung Transplantation (ISHLT) Consensus Document on MCS Infections recommends intravenous antibiotics to target Staphylococcus species administered within 1 hour of skin incision. Among those whose nares are positive for methicillin-resistant Staphylococcus aureus (MRSA) and in hospitals with a high prevalence of MRSA, the recommendation is vancomycin administered intravenously at least 2 hours prior to skin incision. Current recommendations do not advocate broad-spectrum gram-negative or antifungal prophylaxis; however, this depends upon the institutional microbiological data and resistance patterns. Equally important is the meticulous management and care of the percutaneous driveline. Management of the driveline in the early postimplantation period primarily includes placement of a protective dressing, a stabilizer, and daily wound cleansing with a mild antiseptic until there is no longer drainage and the wound appears to have superficially epithelialized after which time chlorhexidine is utilized one to three times per week with dressing changes, under sterile conditions with driveline immobilization ( Fig. 12.2 ). Our practice at University of Alabama at Birmingham is to use two driveline fixation devices in addition to a suture at the exit site. The use of a stabilizing percutaneous exit site suture is controversial. If the surgeons place an exit site suture, it is important to remove it within 2–3 weeks of implantation to prevent accumulation of detritus around the stabilizing suture, which encourages infection.




Fig. 12.2


Demonstration of a commercially available driveline dressing kit including dressing material, skin cleanser, a “biopatch,” and stabilizer/tension relief.

(© Medline Industries, Inc. 2019.)


Prevention of driveline trauma begins when the patient is mobilized. It is important to assign a sufficient number of helpers to move the patient from a sitting position to a standing one and during ambulation early after surgery. Dropping the controller can pull the driveline fixation devices away from the skin and traumatize the exit site during the initial phase of healing, thereby predisposing the wound to infection. Excellent driveline care using standardized protocols for driveline implantation and management can decrease the incidence of driveline infection, which remains at about 20% in the first 12 months after device implantation. If a patient drops the controller or in any other way traumatizes the driveline exit site, the patient will require careful initial inspection of the driveline exit site and frequent monitoring, as trauma to the exit site is one of the leading causes of bacterial driveline infections. Directly asking the patient about any instances of driveline exit site trauma at each postdischarge clinic visit should be a routine part of the history.


The cumulative incidence of driveline infections is high and may represent the Achilles heel of pumps that require percutaneous energy delivery. Several groups are working on novel transcutaneous energy transmission systems, and the first generation of fully implanted pumps (i.e., pumps with no percutaneous components) has been tested. Patients and physicians agree that fully implanted systems are highly desirable. However, there are several important design challenges. The implanted components (energy receiver, controller, and back-up battery) must be highly reliable, not result in additional infection risk, and not be prohibitively expensive to manufacture. The external energy transmitter and receiver coupling may only cause mild local warming and tissue pressure to prevent local skin damage. Replacement of individual components that fail (e.g., controller) or reach end of life (e.g., back-up battery) must have acceptable associated morbidity and hospital length of stay. A fully implantable VAD system is not currently available, although there were prior trials of a fully implanted LVAD and a fully implanted total artificial heart.


Risk factors for driveline infection include obesity, cachexia, and malnutrition. Several methods have been developed for preimplantation nutritional risk assessment. However, it remains challenging to actually improve a patient’s nutritional state prior to VAD implantation or in the early phase of postimplantation recovery due to hemodynamic instability, poor gastrointestinal motility, depressed appetite, and the patient’s catabolic state in severe heart failure or early after surgery. Some have advocated for the initiation of parenteral nutrition prior to implantation and continuing through the postimplantation period to provide uninterrupted protein and calorie support. This has been achieved without a concomitant increase in line-related infection. It is nevertheless important to aggressively optimize the patient’s nutrition before and after implantation.


Inflow cannula position should be checked in the operating room with transesophageal echocardiography after the chest is closed. Typically, this is done with transthoracic echocardiography early after return to the ICU, and then as needed.


Optimal positioning of the inflow cannula relative to the LV walls and the mitral valve orifice is crucial to pump function. The anteroposterior (AP) chest radiograph (CXR) provides a reasonable initial determination of pump position for axial-flow devices (e.g., HeartMate II; Abbott Inc., Abbott Park, IL). Pump position (especially the angle between the pump housing and the inflow cannula for axial flow designs) is one factor that appears to influence pump thrombosis. Moreover, suboptimal inflow cannula position will limit pump flow due to obstruction of the inflow cannula. Excessive outflow graft length or suboptimal positioning of the outflow graft anastomosis on the aorta can lead to kinking that diminishes pump flow and can cause hemolysis. With the initial HeartMate 3 experience (Abbott Laboratories), excessive outflow graft twisting or kinking resulted in a US Food and Drug Administration class I recall due to the frequency of flow interruption and thrombus formation when these events occurred. Positioning of centrifugal LVADs (e.g., HeartWare, Medtronic, Minneapolis, MN; and HeartMate 3, Abbott, Abbott Park, IL) has not been studied as extensively as axial flow pumps. However, an AP CXR will provide an estimate of inflow cannula position relative to the mitral orifice. For both axial and centrifugal flow pumps, echocardiography and three-dimensional reconstructions of gated computed tomography (CT) images provide the most precise information to describe the position of the inflow cannula relative to the ventricular walls and the mitral orifice. Echocardiography has the added advantage of flow assessment within the LV. Continuous laminar flow from the mitral orifice to the inflow cannula suggests an optimal inflow cannula position. Shifting of pump position within the body following implantation has been documented for axial flow devices. These shifts, if sufficiently severe, are associated with increased incidence of pump thrombosis. Shifts of inflow cannula position in centrifugal pumps are less of a problem than in axial flow VADs; however, in small patients and in patients with massive ventricles, the chest wall can push the inflow cannula for a centrifugal pump medially into a position near the chordae tendineae or mitral leaflets. Leaving the pericardium open or adding relaxing incisions to the right and/or left pericardial edges will allow the heart to move away from the pump.


Another important aspect of device management in the early postoperative period is modulation of pump speed (rpm). When the speed of a continuous-flow pump is changed, the results should be assessed not just by pump output but also by the effect on pulsatility (either the Pulsatility Index (PI) for the HeartMate II or 3 device or wave form amplitude on the HeartWare HVAD device). Most speed changes should be accompanied by direct visualization, with careful attention to avoid septal suction toward the left or bowing. Diminished pulsatility or loss of pulsatility suggests that the rpm is set to too high, which predisposes the pump to suction events and can worsen RV contractility. Loss of RV contractility because of septal shifting is an important concept that was developed by David Farrar, John Woodard, and Edna Chou in a canine heart failure model. The ventricular septum provides an important contribution to RV contractile function (up to 20%–40%). When an excessively high pump speed pulls the ventricular septum toward the LV, the septal contribution to RV function is lost. Combined with the resultant RV dilatation and increased wall stress, this can lead to a cycle of diminishing cardiac output (LVAD + native LV) and rising central venous pressure (CVP).


Assessment of the VAD patient’s incision sites, for instance, the sternotomy incision and/or thoracotomy incision, is similar to other surgical procedures. Wounds should be periodically inspected for the development of incisional separation, erythema, or pain. Purulent discharge suggests that a superficial wound infection may be present, while sternal instability may indicate sternal dehiscence or osteomyelitis (mediastinitis). Mediastinitis is a grave complication in a patient with an implanted LVAD and should be treated aggressively following diagnosis. Diagnosis of mediastinitis is based on wound culture, blood culture, wound assessment, leukocyte count, and imaging of the sternum with plain films and CT scan. Serious complications with deep wound infections were encountered early in the use of durable implanted devices. The large space required for pulsatile implanted LVADs predisposed patients to infections of the pump pocket that could involve the inflow and outflow cannulas. Success in managing these patients, at least for limited periods of time to allow transplantation, was achieved with antibiotic-eluting beads of bone cement and the use of autologous tissue such as rectus flaps and omental transfers. Similar methods can be used to manage infections of implanted rotary pumps and mediastinitis in patients with VADs. The collaboration of plastic-reconstruction surgeons to plan and perform these operations is extremely helpful, especially when components of the pump are exposed. Fortunately, deep infections that involve the pump or cannulas are less common with rotary pumps because they require a smaller pocket than a pulsatile LVAD does (e.g., axial flow pumps) or no pocket (e.g., centrifugal pumps attached to the LV apex).




Early postimplantation medical management in the ICU


Postimplantation, patients will arrive intubated and sedated to the ICU with chest tubes, a mediastinal tube, a pulmonary artery catheter, an arterial line, an orogastric tube, peripheral intravenous lines, and, generally, at least one inotrope and one vasopressor running. Vasoplegic patients, or those with RV failure, may be receiving high doses of multiple inotropes and vasopressors, and they may have inhaled nitric oxide or epoprostenol connected to the ventilator. Some may have a temporary RVAD, surgical or percutaneous, in place.


Initial attention should be paid to transferring the many drips from anesthesia pumps to the ICU pumps, without interrupting the infusions (or for the shortest time possible), leveling the arterial-line and pulmonary artery catheter (PAC) transducers so the hemodynamic monitoring is accurate, and placing the patient on the ICU ventilator. Often, during this period of transition, the patient’s stability may be tenuous at best and several nurses, an intensivist, respiratory therapy, and a medical assistant should be on hand to receive the patient. Fluid shifts during this period are expected, including those from bleeding and third spacing due to inflammation and capillary leak. These events may require both resuscitation with crystalloid, colloid, or blood and increased vasopressor or inotropic support for vasoplegia and RV failure, respectively. All ICU intravenous infusions should be prepared in advance, including inotropes, vasopressors, and sedatives. Crystalloid and colloid should be immediately available should the blood pressure become dangerously low. The patient’s mean arterial pressure (MAP) on the arterial line should be maintained at > 65 mm Hg and no higher than 80 mm Hg. The amount of blood drained into the chest/mediastinal tube containers should be marked with time and date. Finally, a trained LVAD nurse or coordinator should inspect the driveline dressing to be sure it has been placed correctly and the tension-relief stabilizer is likewise being used and appropriately positioned.


Echocardiographic and hemodynamic speed optimization is an iterative process, initiated in the operating room and performed as needed in the ICU and prior to discharge with the same basic targets in mind. The LVAD speed is initially optimized in the operating room with transesophageal echocardiogram (TEE) guidance. It should be at the lowest speed possible, which generates adequate flow (confidence interval [CI], 2.2–2.6 L/Min × M 2 ), does not suck the septum leftward, minimizes mitral regurgitation (MR), and allows a perfusing blood pressure (often with the support of a low-dose vasopressor infusion for improved vascular tone). The reason for minimizing the pump speed and LV output from the pump is to avoid an acute volume load on the RV since it is not accustomed to robust venous return, is often dysfunctional as well, and without the benefit of mechanical support. Rather the speed should be increased slowly over the first few hours to days and likely will need to be increased as third-spaced fluid returns to the intravascular space and the LV needs increased unloading. Evidence suggests that after this initial perioperative phase, pump speed changes among patients with an implanted LVAD are much less likely to affect the RV performance if done carefully.


If the patient arrived in the operating room with an intraaortic balloon pump (IABP) in place, the augmentation is decreased and the IABP is removed as soon as possible after arrival in the ICU. The pulsation of the IABP creates a pulsatile waveform; however, conventional wisdom has it that IABP inflation will stop the forward flow or actually reverse the flow of blood out of a rotary LVAD. On the other hand, a small study in 2015 at Barnes-Jewish Hospital in St. Louis looked at 51 flow tracings from LVAD patients with an IABP in situ postimplantation at inflation ratios of 1:1, 1:2, and 1:3 versus off and actually found no statistically significant difference in mean flow among groups, although there was a statistically significant reduction in minimum flow rates and increase in maximal flow rates, thus widening the pulsatility. However, since the IABP is often on a 1:2 or 1:3 setting postimplantation, there is a risk of thrombus formation and thromboembolism from the balloon pump. As long as the activated clotting time (ACT) has dropped sufficiently, the balloon pump should be pulled, and appropriate hemostasis achieved.


The initial hours in the ICU are critical to establishing stability and recognizing potential problems that, if untreated, may lead to increased morbidity, mortality, and prolonged ICU length of stay. These include bleeding, progressive right heart dysfunction, vasoplegia, and pericardial tamponade. Close attention to vital signs, invasive hemodynamics, mixed venous oxygen saturation, and serum lactate values will often hint at potential problems ahead. One must also be attuned to changes in VAD parameters, urine output, and chest tube output.


The development of a low pump output is a frequent clinical challenge in the early postoperative period. Clinicians must pay particular attention to trends in the CVP levels as well as the hourly chest tube output and LVAD parameters such as pump flow and pulsatility. The postoperative events that lead to low pump output include bleeding, hypovolemia, pericardial tamponade, and acute RHF ( Fig. 12.3 ). Differentiating one from the other can be accomplished by using invasive hemodynamics and clinical acumen. For example, in the case of tamponade, an elevation of CVP and the sudden cessation of mediastinal tube drainage or an insidious fall in pump output should prompt echocardiographic confirmation and early chest exploration. Other typical signs, such as muffled heart sounds and paradoxical pulse, are unreliable in the setting of a continuous-flow blood pump and a recent sternotomy. Acute RHF can be identified by a sudden drop in flow, a dampening of wave forms on a centrifugal-flow device with a waveform display, and/or a sudden drop in the mixed venous oxygen saturation combined with a rising CVP and serum lactate level. A transesophageal echocardiogram can help rapidly make the diagnosis and the surgeon may intervene in time (by placing a temporary RVAD) to prevent damage to end-organs from low cardiac output ( Fig. 12.4 ).


Dec 29, 2019 | Posted by in CARDIOLOGY | Comments Off on Postoperative VAD Management: Operating Room to Discharge and Beyond: Surgical and Medical Considerations

Full access? Get Clinical Tree

Get Clinical Tree app for offline access