Assessing Thromboembolic Risk on Anticoagulation Reversal in Patients with Left-Ventricular Assist Devices

In this issue of the American Society for Artificial Internal Organs journal, Jennings et al.1 report results of a retrospective study assessing the thromboembolic (TE) risk on reversal of anticoagulation in patients with continuous-flow left-ventricular assist devices (CF-LVADs). In this study, 25 patients among a cohort of 122 patients with CF-LVAD experienced 38 anticoagulation reversal events. The rate of thromboembolism within 30 days of attempted reversal was 2.6% and the mortality rate within 30 days of reversal was 20%. Since the publication of the Randomized Evaluation of Mechanical Assistance for the Treatment of Congestive Heart Failure trial in 2001,2 significant progress has been made in the field of mechanical circulatory support. The current-generation CF-LVADs are less bulky, more durable, and have a significantly longer battery life. Patient selection, surgical techniques, and postimplantation clinical management continue to evolve and be refined, reducing the frequency of device complications and thereby patient morbidity and mortality. Although there have been improvements made in the blood–LVAD interface, the problem still persists. The nonphysiological shear stress imparted on blood components as they pass through the LVAD and other potential mechanisms of bleeding in patients with CF-LVADs are neatly outlined in a review by Suarez et al.3 Quoting the Sorites paradox “Removing a single grain does not turn a heap into a nonheap, but if the process were repeated enough times, is a single remaining grain still a heap. If not, when did it change from a heap to a nonheap?” Similarly in each individual patient with a CF-LVAD, how much anticoagulation is adequate enough to prevent device thrombosis and TE events yet prevent bleeding complications is unknown. It has been suggested, and is reasonable to assume from our experience in patients with atrial fibrillation, that patients tend to bleed more when the INR is greater than 3, whereas patients with INR less than 1.5 tend to have higher TE event rates. In trying to determine this “ideal anticoagulation range or optimal INR threshold,” one should keep in mind that each patient’s inherent bleeding or TE risk is different. To our knowledge, there is no definitive method that one could apply to identify which patient will be more at risk of developing a gastrointestinal arterio-venous malformation and thereby a bleed or an intracranial bleed post–CF-LVAD. Probably, one of the best recent studies to examine predictive factors for bleeding and TE events in a cohort with CF-LVADs is from Boyle et al.,4 who demonstrated that age, basal metabolic index, sex, baseline hematocrit, and etiology of heart failure are significant predictors of bleeding or TE events. Based on the lack of conclusive evidence and the tremendous variability in patient profiles, current mechanical circulatory support guidelines5 published by the International Society of Heart and Lung Transplantation do not make any definitive recommendations on how to approach this issue. More often than not, healthcare providers use their clinical gestalt and take into consideration factors that increase risk for bleeding or TE events, such as patient’s age, prior history of systemic bleed, renal function, and atrial fibrillation history, and attempt to individualize the anticoagulation goal for the specific patient. Despite being meticulous about INR management, it is well known from prior anticoagulation studies that INR levels are extremely dynamic and the time spent in therapeutic range is often less than 60%.5 Hence, we must be cognizant of the fact that some patients will have bleeding events on anticoagulants, no matter how good we are at managing their anticoagulation and thereby will require discontinuation of their anticoagulants. The authors should be commended for bringing attention to the topic of TE risk on anticoagulant discontinuation, as there are currently no uniformly accepted or researched anticoagulation reversal strategies in this patient cohort. In this study, 25 of the 122 patients studied received anticoagulation reversal for clinical scenarios that are commonly encountered in this cohort, namely: acute hemorrhage secondary to gastrointestinal bleed, intracranial bleeding, and miscellaneous bleeding for invasive procedures, increased INR values, and unexplained anemia. What the investigators have done nicely is to inform practioners and to provide readers with a contemporary descriptive analysis of TE risk on discontinuing anticoagulation in patients with CF-LVADs. The study does not offer an INR cut-off at which to reverse anticoagulation and rightly so. Because the incidence of TE event risk between different reversal agents such as fresh-frozen plasma (FFP) or vitamin K is unknown, the authors also attempted to examine event rates between reversal strategies (FFP vs. vitamin K). But as the authors acknowledge, patients were screened for this study based on receipt of vitamin K; hence, patients reversed with FFP were not included in this study and thereby event rates between the two cannot be accurately assessed. Routine use of vitamin K in the preoperative period before LVAD placement has been shown to reduce postoperative bleeding and the need for surgical re-exploration after implantation of pulsatile LVAD, but its current practice has been debated in CF-LVADs as it may complicate postoperative anticoagulation management.6 Reversing warfarin-induced anticoagulation by vitamin K can be time consuming when compared with FFP and prothrombin complex concentrate. The administration of FFP is not ideal, as it requires substantial intravenous volume which is not appropriate in patients with heart failure. Prothrombin complex concentrate, which contains a high concentration of the vitamin K-dependent coagulation factors II, VII, IX, and X, has been reported to be effective for rapid reversal of warfarin-induced anticoagulation.8 Prothrombin complex concentrate has been reported to reduce the PT-INR faster than FFP or vitamin K.9,10 Although the authors suggest that the TE event risk on reversal of anticoagulation is low, it is important to note the limitations of the data presented. The small number of patients always presents a challenge to make any strong conclusions. Diagnostic assessment to identify potential subclinical pump thrombosis was not performed (either by measuring lactate dehydrogenase or RAMP echocardiograms)11; hence, the reported 30 days of TE risk event rate of 2.6% may be falsely low. Furthermore, the authors report 30 days of TE event rates, but if a patient developed device thrombus during the period of anticoagulation reversal or during resumption of anticoagulation, it could potentially present as a complication much later. Because this is a retrospective study, it is impossible to know whether the intervention made (INR threshold at which anticoagulation was held or continued) affected patient outcomes in a positive or negative way. As with most studies in this cohort, this study inspires more questions. Some important unanswered questions include the following: 1) What INR value equates to sufficient anticoagulation reversal without significantly increasing TE events, 2) Once we reverse anticoagulation, how long do you maintain a subtherapeutic anticoagulation profile before uptitrating therapy, 3) How or should we incorporate patient’s hemoglobin level in determining the INR goal, 4) What agent should be used to reverse anticoagulation in this cohort, and 5) What are the bleeding to TE event rates in patients who take longer to achieve therapeutic anticoagulation on resumption of anticoagulant therapy. The challenge is in deciphering what each patient’s risk for TE event vs. pump thrombosis is, and additional research is needed to discern the specific criteria for risk stratification. Multicenter studies among mechanical circulatory support programs and a strong collaborative effort with our hematology, gastroenterology, and neurology colleagues are urgently needed to improve our understanding of this complex clinical topic.