Blood transfusions in ICU: double-edged sword Paul Hébert, MD MHSc(Epid) Physician-in-Chief, CHUM Professor, University of Montreal
Canadian Critical Care Trials Group Collaborating for Impact Leading Science, Saving Lives www.ccctg.ca May 2014
Objectives Review evidence on transfusion triggers Describe upcoming studies on transfusion triggers MINT
Purpose: Are a restrictive and liberal red cell transfusion strategy are equivalent in terms of effects on mortality and morbidity in critically ill patients Hebert et al. NEJM 321: 151-156, 1999
Survival of all patients over 30 days 100 Survival (%) 90 80 70 Restrictive strategy Liberal strategy 18.7% 23.3% 60 p=0.10 50 0 5 10 15 20 25 30 Time (Days) Hebert et al. NEJM 321: 151-156, 1999
TRIPICU - Outcomes Restrictive group Liberal group Total # of patients 320 317 Non-transfused (n) 174 (54.4%) 7 (2.2%) No. of transfusions 301 542 New or progressive MODS (n) 38 39 New or progressive MODS (%) 11.9 (95% CI 8.5-15.9) 12.3 (95% CI 8.9-16.4) Lacroix, Hebert et al. NEJM 2007; 356: 1609.
Cardiovascular Disease Patients with cardiovascular diseases (n=357) Patients with ischemic heart disease (n=257) 100 100 Survival (%) 90 80 70 60 Liberal p = 0.95 Restrictiv Survival (%) 90 80 70 60 Liberal p = 0.30 Restrictive 50 0 5 10 15 20 25 30 Time (Days) 50 0 5 10 15 20 25 30 Time (Days) Hebert et al. NEJM 321: 151-156, 1999
Complications Complication Liberal Restrictive P Values (n=420) (n=418) Cardiac No. (%) 88 (21.0) 55 (13.2) <0.01 Myocardial Infarction 12 (2.9) 3 (0.7) 0.02 Pulmonary Edema 45 (10.7) 22 (5.3) <0.01 Angina 9 (2.1) 5 (1.2) 0.28 Cardiac Arrest 33 (7.9) 29 (6.9) 0.6 Pulmonary No. (%) 122 (29.1) 106 (25.4) 0.22 ARDS 48 (11.4) 32 (7.7) 0.06 Pneumonia 86 (20.5) 87 (20.8) 0.92
Effect of anemia on mortality in cardiac disease Odds Ratio 16 13 10 7 Healthy (No IHD) Ischemic Heart Disease P=0.03 Retrospective cohort of patients who refuse blood transfusion CVD definition - History of MI, angina, CHF, or PVD. 1,958 patients age 18 or older. Undergo surgical procedure in OR. Outcome-30-day mortality or morbidity. 4 1 6 7 8 9 10 11 12+ Preoperative Hgb (g/dl) Carson JL, et al. Lancet 1996;348:1055-60.
Copyright restrictions may apply. Rao, S. V. et al. JAMA 2004;292:1555-1562.
Risk of Death in 24,112 transfused versus non-transfused from 3 RCTs Nadir Hematocrit, % adjusted OR (95% CI) 35 291.64 10.28-8273.85 30 168.64 7.49-3797.69 25 1.13 0.70-1.82 20 1.59 0.95-2.66 Rao, S. V. et al. JAMA 2004;292:1555-1562.
Hebert and Fergusson,JAMA, 2004 How do Rao and Wu studies compare? Both studies document harm from RBC transfusion with hematocrits exceeding 33% Reasons for differences at hematocrits< 33% Different population (younger and aggressively treated in Rao study) Different data acquisition Primary data collection vs administrative database Different statistical techniques Different event rates
Functional outcomes in cardiovascular patients undergoing surgical hip fracture repair (FOCUS) Design: Multicentre RCT in 47 North American centres Study Population: 2016 hip fracture patients undergoing surgical repair with a Hb < 100 g/l within 3 days fo surgery Intervention: Liberal Strategy: transfusion trigger of 100 g/l Restrictive Strategy: transfusion for symptomatic anemia Outcomes: Primary: functional recovery (ability to walk 10 feet without human assistance 60 days post-op) Long term survival, nursing home placement, post-op complications (MI and infection) Carson et al. NEJM 2011, pp2453-62.
Carson JL et al. N Engl J Med 2011;365:2453-2462 Lowest Daily Hemoglobin Levels
Hospital Outcomes Carson JL et al. N Engl J Med 2011;365:2453-2462
Death, MI or Unscheduled Revascularization at 30 days, *p=0.054, adjusted for age p=0.076
From: Transfusion Requirements After Cardiac Surgery: The TRACS Randomized Controlled Trial JAMA. 2010;304(14):1559-1567. doi:10.1001/jama.2010.1446 Figure Legend: Time zero was just after randomization (12 hours before surgery). Hazard ratio, 1.28 (95% confidence interval, 0.60-2.73) (P =.99) for restrictive strategy vs liberal strategy. Date of download: 11/1/2014 Copyright 2014 American Medical Association. All rights reserved.
Study overview Design: Multicentre RCT Study Population: 7800 patients with a myocardial infarction and hemoglobin below 100 g/l Intervention: Liberal Strategy: transfusion trigger of 90 g/l Restrictive Strategy: 80 g/l with transfusion 1 unit at a time Outcomes: Primary: composite of mortality and myocardial infarction at 30 days Secondary: acute MI, unscheduled revascularization, CHF and readmission
Recommendations Adopt a transfusion threshold between 70 g/l and 80 g/l in patients with acute coronary syndrome.really not sure Cardiac surgery.studies completed and underway Evidence also supports a lower transfusion threshold in: Sepsis and septic shock Acute GI bleeding Perioperative care
Recommendations Join the MINT trial If you want to know what to do with acute coronary syndromes!