Comparative Effectiveness of Different Strategies for Monitoring Neonates on Extra Corporeal Membrane Oxygenation (ECMO)

Background: ECMO can be life-savingin patients with cardiac or respiratory failure. Anticoagulation, typically with unfractionated heparin (uFH), is necessary to prevent the ECMO circuit from clotting. However, titrating the intensity of anticoagulation to maintain the balance between preventing thrombotic and hemorrhagic complications remains a challenge for providers. This is particularly true in neonates who are at the highest risk of intracranial hemorrhage, and in whom titrating uFH is notoriously difficult. In order to improve titration of anticoagulation, we instituted Enhanced Anticoagulation Monitoring Guidelines for uFH in the neonatal intensive care unit (NICU) at the Children's Hospital of Philadelphia (CHOP) in May of 2012. The guidelines included additional laboratory studies including uFH anti-Xa levels, antithrombin III levels, and aPTT compared to prior practice which relied mainly on ACT alone. The guidelines also included recommendations for blood product transfusion for neonates with dilutional coagulopathy (prolonged ACT or aPTTwith a low uFH anti-Xa level). In this comparative effectiveness study, we evaluated how the enhanced guidelines influenced care compared to prior practice. Here, we report the difference in thrombotic and hemorrhagic complications as well as in blood product utilization. Ongoing analysis includes heparin exposure and cost analysis of the enhanced guidelines. Methods:Patients in the CHOP NICU treated with ECMO from 2009-2014 were included in this retrospective study. Data collection included demographics, ECMO indication, thrombotic or hemorrhagic findings on head imaging, and thrombotic or hemorrhagic complications (including circuit changes) while on the ECMO circuit. Intracranial hemorrhage was defined as consistent mention of intracranial bleeding on two or more consecutive head ultrasounds while a patient was on the ECMO circuit. In addition, transfusion data (red blood cells, fresh frozen plasma, and platelets) was obtained. These initial values (in mL) were adjusted by patient weight in kilograms and hours on the ECMO circuit to allow for utilization comparisons in mean mL/kg/hour. Analysis comparing relative frequencies of significant bleeding and clotting outcomes was then conducted via Fisher's exact tests. Lastly, Student's t-tests of independent groups were conducted to compare the mean utilization rates of blood products before and after the institution of the guidelines. Results: A total of 127 neonates treated with ECMO during the study period were identified (62 patients before the guidelines and 65 patients after). The distribution of ECMO indications per group are listed in table 1 and were not significantly different (p=.112). There was a similar incidence in circuit change events between the two groups. While the frequency of acute CNS ischemic events decreased from 14.5% to 9.2%, this was not statistically significant (p=0.36). With respect to bleeding events, the pre-intervention group had a 37.1% incidence of intracranial hemorrhage per head imaging compared to 24.6% in the post-intervention group though this difference was not of statistical significance (p=.127). While the rates of utilization of packed red blood cells did not differ between the two groups (p=0.76) suggesting total hemorrhagic complications may not differ, the use of both fresh frozen plasma and platelets did increase significantly after initiation of the enhanced guidelines (p=0.02 and p<0.005, respectively). Of the cohort as a whole, 2 patients required immediate discontinuation of the ECMO circuit due to evidence of ischemic stroke with hemorrhagic conversion on imaging but survived. A total of 7 patients died secondary to ECMO-induced bleeding or clotting complications (intracranial hemorrhage (n=3), postoperative bleeding (n=1), ischemic stroke (n=2), pericardial tamponade (n=1)). Conclusions: Changes in clinical practice are often made using best available evidence to improve patient care. Therefore, when these changes are implemented, it is critical to evaluate their impact. In our initial data analysis, we did not identify significant decreases in bleeding or clotting complications with initiation of enhanced anticoagulation guidelines. Additional analysis (heparin usage, cost analysis, and clinician's satisfaction with the enhanced guidelines) is underway to fully understand the impact of the changes. Disclosures No relevant conflicts of interest to declare.