Objectives: To compare hemoglobin concentration (Hb), RBC use, and patient outcomes when restrictive or liberal blood transfusion strategies are used to treat anemic (Hb = 90 g/L) critically ill patients of age = 55 years requiring = 4 days of mechanical ventilation in ICU. Design: Parallel-group randomized multicenter pilot trial. Setting: Six ICUs in the United Kingdom participated between August 2009 and December 2010. Patients: One hundred patients (51 restrictive and 49 liberal groups). Interventions: Patients were randomized to a restrictive (Hb trigger, 70 g/L; target, 71-90 g/L) or liberal (90 g/L; target, 91-110 g/L) transfusion strategy for 14 days or the remainder of ICU stay, whichever was longest. Measurements and Main Results: Baseline comorbidity rates and illness severity were high, notably for ischemic heart disease (32%). The Hb difference among groups was 13.8 g/L (95% CI, 11.5-16.0 g/L); p <0.0001); mean Hb during intervention was 81.9 (SD, 5.1) versus 95.7 (6.3) g/L; 21.6% fewer patients in the restrictive group were transfused postrandomization (p <0.001) and received a median 1 (95% CI, 1-2; p = 0.002) fewer RBC units. Protocol compliance was high. No major differences in organ dysfunction, duration of ventilation, infections, or cardiovascular complications were observed during intensive care and hospital follow-up. Mortality at 180 days postrandomization trended toward higher rates in the liberal group (55%) than in the restrictive group (37%); relative risk was 0.68 (95% CI, 0.44-1.05; p = 0.073). This trend remained in a survival model adjusted for age, gender, ischemic heart disease, Acute Physiology and Chronic Health Evaluation II score, and total non-neurologic Sequential Organ Failure Assessment score at baseline (hazard ratio, 0.54 [95% CI, 0.28-1.03]; p = 0.061). Conclusions: A large trial of transfusion strategies in older mechanically ventilated patients is feasible. This pilot trial found a nonsignificant trend toward lower mortality with restrictive transfusion practice.