The literature is full of observational studies that show that blood is harmful. A recent systematic review by Malik and Corwin (Crit Care Med 2008;36:2667-74) found that in 42 of 45 observational studies he risk of transfusion outweighed the benefit. Unfortunately, it is not possible to fully control for confounding because blood transfusion is a marker for severity of disease. Only a randomized clinical trial can evaluate the risk vs. benefit of transfusion in an unbiased manner.
– Jeffrey L. Carson