Adverse Events and Mitigation Strategies

Introduction

The evolution of ventricular assist devices (VAD) since the year 2000 has allowed patients with end-stage heart failure to survive with increased functional capacity and quality of life (QOL). Since the beginning of randomized clinical trials in the field, survival, QOL, and resource requirements have been limited by adverse events. In the case of the Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure (REMATCH) trial, pump malfunction due to nondurable components requiring a surgical exchange and stroke caused significant morbidity and mortality. As engineering in the field of mechanical circulatory support (MCS) has applied novel solutions to increase durability and to miniaturize the pumps for more facile implantation, the incidence and manifestations of certain adverse events have been altered or amplified, such as gastrointestinal (GI) bleeding and neurologic dysfunction, with a more complex pathogenesis in association with continuous-flow (CF) device support. As pump technology evolves over time, we should expect that newer devices designed to improve upon previous platforms will be associated with adverse events that require rigorous investigation and novel proposals to eliminate or mitigate their occurrence.

MCS with current-generation CF devices has provided a major improvement in duration of survival on device compared to prior pulsatile pumps. With current documented survival exceeding 80% at 1 year and 70% at 2 years, extension of this therapy into the ambulatory advanced heart failure population (Interagency Registry for Mechanically Assisted Circulatory Support [INTERMACS] levels 4–6) has been extensively discussed and implemented to a limited degree. However, the success of this therapy is a function not only of survival but also of QOL among survivors. A major determinant of QOL with CF pumps is the frequency, severity, and sequelae of major adverse events. Grady and colleagues demonstrated the major deleterious effect on QOL of certain adverse events at 6 months postimplantation. The Randomized Evaluation of VAD Intervention Before Inotropic Therapy (REVIVE-IT) trial was designed to test the hypothesis that device therapy was superior to optimal medical therapy for such patients. However, the presence of a higher-than-expected incidence of pump thrombosis in the HeartMate II (HMII) device during the era of the planned study caused the investigators to terminate the study prior to performing the first implantation. With increased awareness of the frequency and impact of major adverse events during CF MCS support, the application of these devices in patients with ambulatory class IV advanced heart failure has declined during recent years. The proportion of patients in INTERMACS levels 4–7 at implantation represented 18% from 2012 to 2014 and decreased to 13% in 2015–2016. With the exception of right heart failure, other adverse events are equally as likely in ambulatory heart failure patients as in patients more critically ill ( Figs. 13.1–13.3 ). Thus, understanding the frequency, severity, and impact of major adverse events is of critical importance; both for truly informed decision making by patient, family, and caregiver and to inform decisions about future pump design and patient management. Moreover, increased device durability and dramatically lower adverse events (AE) rates must be achieved before a substantial increase in mechanical circulatory support device (MCSD) usage occurs in less ill populations (e.g., INTERMACS levels 4–7).

Definitions

Accurate identification of adverse events and reliable calculation of incidence requires clear definitions. Throughout the history of MCS trials, differing applications and definitions have been applied. When INTERMACS was created in 2006 as the national database for US durable MCS devices approved by the Food and Drug Administration, extensive deliberations among expert clinicians, industry, governmental personnel, and engineers focused on adverse event definitions. These were incorporated into the national INTERMACS database. More recently, the Academic Research Consortium has participated in attempts to harmonize definitions of adverse events among countries. The definitions used by INTERMACS in the collection of postimplantation data on more than 20,000 patients are included in the Appendix .

Time-related occurrence, contributing factors, and causation

Adverse events occurring early after implantation are often ascribed to the preoperative condition of the patient and often reflect the risks associated with preoperative morbidity—major cardiac surgery is being performed in patients with advanced heart failure, often decompensated and with altered metabolic status and end-organ injury. In contrast, the more chronically occurring events generally result from device design issues or management strategies.

From an analytic perspective, both the time to occurrence of a given adverse event and its repetitive nature (incidence) help characterize its impact. Although the exact “cutoff” is rather arbitrary, multiple analyses have identified different adverse events and frequencies that are more prominent early or later after implantation ( Table 13.1 ). Additional challenges are highlighted by the following examples. A single stroke can have relatively minor or devastating effects on the patient’s QOL. Strokes are discrete events in many cases, although extension of a stroke by hemorrhagic conversion may occur. By contrast, infections and especially device-related infections have slow onset so that the date of an infection AE often describes the date of diagnosis rather than the date of onset. Moreover, the date for resolution for an infection may reflect the cessation of antimicrobial therapy rather than a bona-fide eradication of infection. For example, driveline percutaneous exit site infections are typically treated for weeks with antibiotics and then the antibiotics are stopped. If erythema and purulent drainage at the site worsen, the patient will presumably be placed back on antibiotic therapy. Should this be counted as a de novo infection or persistence of a prior infection? If a driveline percutaneous exit site infection occurs, should all other episodes of driveline infection be considered recurrences of the initial event or only if the infecting organism changes? If the patient is placed on chronic suppressive antibiotic therapy, should the initial episode of driveline infection be considered as resolved or ongoing? Thus, considerable analytic challenges exist as we seek “truths” about the potential life-threatening complications of this therapy.

Adverse event burden and era effect

The overall impact of adverse events can be discussed under the descriptive category of “adverse event burden.” While nonquantitative, depictions such as freedom from any of a group of adverse events are a first step in such a description ( Fig. 13.4 ). As experience has accumulated in the care of patients with CF devices, the incidence of many adverse events has gradually decreased ( Table 13.2 ). This era effect is also in part accounted for by cumulative experience with these devices over time. For example, evaluation of multiple adverse events, along with survival (event-free survival), was an integral part of clinical trials with the HeartMate 3 (HM3) magnetically levitated centrifugal flow pump with wide pathways engineered to minimize friction during propulsion and avoid stasis with an intrinsic pulse design. The primary endpoint of the Multicenter Study of MagLev Technology in Patients Undergoing Mechanical Circulatory Support Therapy with HeartMate 3 (MOMENTUM 3) trial was survival free of disabling stroke and pump exchange. This combined endpoint showed superiority in the HM3 versus the HMII at 6 months and 2 years, largely related to a major reduction in pump exchange for thrombosis. The primary endpoint of comparing the CE Mark trial for the HM3 device to a recent report of “real-world” experience in Germany demonstrated improvement in the incidence of multiple adverse events during the first year in the postclinical trial experience: GI bleeding (12%–7%), right heart failure (10%–7.4%), stroke (18%–0%), and all bleeding (44%–30%).

Adverse events contributing to death

The risk of death caused by various adverse events changes over time ( Fig. 13.5 ). The risk of death from right heart failure is greatest early postimplantation, whereas infection death has an early peak and then gradually increases again after about 2 years. The risk of neurologic death decreases rapidly during the first 6 months and then remains relatively constant. When viewed over a 6-year period, neurologic events account for the greatest cumulative risk of death, followed by infection adverse events ( Fig. 13.6 ). The dramatic impact of adverse events on subsequent mortality is underscored in an analysis from INTERMACS, which demonstrates the strong predictive value of multiple adverse events during the first 3 months on subsequent survival ( Table 13.3 , Fig. 13.7 ).

Perioperative bleeding

Perioperative bleeding, with the associated transfusion requirement, increases the risk of infection, allosensitization, and acute lung injury. In the presence of lung injury and destabilizing bleeding, right ventricular failure may lead to reduced left VAD (LVAD) output, end-organ ischemia, and reduced survival. The potential need for reoperation to evacuate clot, particularly if tamponade is present, further exacerbates the spiral of hemodynamic instability and the potential for mediastinal infection. Nearly one-third of patients require reoperation for bleeding post VAD implantation. Several factors increase the risk of an increased bleeding diathesis following VAD implantation. A chronic coagulopathy often accompanies chronic right heart failure. The preoperative Model for End-Stage Liver Disease (MELD) score, which is a weighted sum of serum creatinine, bilirubin, and international normalized ratio (INR), is correlated with the risk of bleeding and death after LVAD implantation. Chronic administration of platelet inhibitors and warfarin add to the coagulopathic state, particularly in the presence of prior cardiac operations. Therefore, an elevated INR should be normalized before implantation. Patients more critically ill at the time of implantation (INTERMACS level 1) are also poised for greater coagulopathy.

Accurate dissection, meticulous hemostasis, reversal of coagulopathy, and careful inspection of suture lines are the key elements of prevention of postoperative bleeding. In the event of a severe ongoing coagulopathy after reversal of protamine following cardiopulmonary bypass, initial placement of a wound vacu-system, followed by secondary sternal closure, is safe and common practice. An irrigation system (e.g., Pulseavac, Zimmer Biomet, Warsaw, IN) used with sterile saline, possibly with 3–5 mL Betadine soap (Purdue Pharma LP, Stamford, CT) added per liter, can be used to clean the mediastinal surfaces after removal of the packing sponges. Care must be taken not to stimulate arrhythmias or excessively cool the heart during the irrigation. The goal is to minimize the residual clot and fibrin deposits in the wound prior to final closure.