OBJECTIVES: To determine whether randomization of patients undergoing extracorporeal membrane oxygenation to either therapeutic or a low-dose anticoagulation protocol results in a difference in activated partial thromboplastin time and anti-Xa. DESIGN: Randomized, controlled, unblinded study. SETTING: Two ICUs of two university hospitals. PATIENTS: Patients admitted to the ICU, who required extracorporeal membrane oxygenation (venovenous or venoarterial) and who did not have a preexisting indication for therapeutic anticoagulation. INTERVENTIONS: Therapeutic anticoagulation with heparin (target activated partial thromboplastin time between 50 and 70 s) or lower dose heparin (up to 12,000 U/24 hr aiming for activated partial thromboplastin time < 45 s). MEASUREMENTS AND MAIN RESULTS: Thirty-two patients were randomized into two study groups that were not significantly different in demographics and extracorporeal membrane oxygenation characteristics. There was a significant difference in the daily geometric mean heparin dose (11,742 U [95% CI, 8,601-16,031 U] vs 20,710 U [95% CI, 15,343-27,954 U]; p = 0.004), daily geometric mean activated partial thromboplastin time (48.1 s [95% CI, 43.5-53.2 s] vs 55.5 s [95% CI, 50.4-61.2 s]; p = 0.04), and daily geometric mean anti-Xa (0.11 international units/mL [95% CI, 0.07-0.18] vs 0.27 [95% CI, 0.17-0.42]; p = 0.01). We found similar results when considering only venovenous extracorporeal membrane oxygenation episodes; however, no difference in daily geometric mean activated partial thromboplastin time between groups when considering only venoarterial extracorporeal membrane oxygenation episodes. CONCLUSIONS: Allocating patients on extracorporeal membrane oxygenation to two different anticoagulation protocols led to a significant difference in mean daily activated partial thromboplastin time and anti-Xa levels between groups. When considering subgroups analyses, these results were consistent in patients on venovenous extracorporeal membrane oxygenation. Our results support the feasibility of a larger trial in patients undergoing venovenous extracorporeal membrane oxygenation to compare different anticoagulation protocols; however, this study does not provide evidence on the optimal anticoagulation protocol for patients undergoing extracorporeal membrane oxygenation.