Coronary Heart Disease and Anemia: To Transfuse or Not?

Abstract & Commentary

Synopsis: A retrospective study of 24,112 patients with coronary heart disease showed increased 30-day mortality in patients who received one or more blood transfusions for anemia compared to comparable patients not transfused.

Source: Rao SV, et al. JAMA. 2004;292:1555-1562.

Rao and colleagues from the Duke Clinical Research Institute analyzed data from 3 multicenter international trials of patients with acute coronary syndromes. In this study, 2401 of 24,112 patients received at least one blood transfusion during their admission for anemia. These patients were compared with comparable patients not receiving a transfusion. Among the transfused patients with hematocrits above 25%, there was a 3.94 times higher risk of death within 30 days. Most of the patients were elderly (average age, 69). Rao et al compared their results with a previous study showing different results,1 and attempted to explain the differences.

In an accompanying editorial,2 Hebert and Fergusson from the University of Ottawa Centre for Transfusion and Critical Care Research analyze these findings based on previous research. The literature is conflicting, but suggest that the findings of Rao et al, do confirm that patients with hematocrits of 30% or higher should not be transfused unless there are other clinical indications.

There is benefit in transfusion for patients with hemoglobins below 7.0 g/dL (hematocrit, 21%). What to do with patients with hemoglobin concentrations of 8-10 g/dL (hematocrit, 24-30%) is unclear. Rao et al recommend judicious use of transfusion, such as giving one unit at a time, and only when there are compelling clinical indications, in this group of patients.

Comment by Joseph E. Scherger, MD, MPH

Elderly patients at risk for heart disease are regularly transfused with packed red cells or whole blood for marked anemia (hematocrits of 30% or less). I was trained never to give just one unit of blood. If blood is indicated, give at least 2 units. This study should bring such behavior to a halt, at least temporarily.

Given the conflicting nature of retrospective observational studies, a randomized controlled trial is called for. Rao et al, use improved analytical methods compared with previous studies, but still their results are not definitive. The decision to transfuse a patient may be based on clinical criteria beyond just the hemoglobin or hematocrit, and such factors may not be fully captured in this analysis. It does appear that transfusion, even if done slowly, may care added risk in patients with acute coronary syndromes.


1. Wu WC, et al. N Engl J Med. 2001;345:1230-1236.

2. Hebert PC, et al. JAMA. 2004;292:1610-1612.