Blood Transfusion in Acute Myocardial Infarction

Abstract & Commentary

Synopsis: Blood transfusion is associated with a lower short-term mortality in elderly patients with acute myocardial infarction and hematocrit < 30%.

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

There is little objective evidence concerning when to transfuse blood in patients with acute myocardial infarction. Thus, Wu and associates examined the Cooperative Cardiovascular Project database of 78,974 medicare beneficiaries aged ³ 65 years who were hospitalized for acute myocardial infarction to assess the prognostic importance of anemia and the effect of transfusion. In this cohort, 43% had a low hematocrit (< 39%), and, in general, they had more comorbidities compared to those with normal hematocrits. Also, they had more complications and progressively higher mortality as the hematocrit fell. Only 5% of the total cohort received a transfusion, and transfusion was more likely the lower the hematocrit. Transfusion was associated with a lower 30-day mortality in patients with hematocrits below 33%, but a higher mortality if the hematocrit was above 36%. Subgroup analysis among patients who survived at least 2 days showed that there was no mortality benefit of transfusion with hematocrit values > 30%. Wu et al concluded that blood transfusion is associated with a lower short-term mortality in elderly patients with acute myocardial infarction and hematocrit < 30%.

Comment by Michael H. Crawford, MD

The most striking finding in this study was the high prevalence of anemia in this elderly, acute myocardial infarction population. Almost half had hematocrits below 39% and one tenth were below 33%. Also, low hematocrit was associated with a higher mortality at 30 days (see Table 1).

Table 1
30-Day Mortality by Hematocrit Range
Hematocrit   5-24% 24-27% 27-30% 30-33% 33-36%  36-39%
Mortality 39% 35%  36% 30%  26% 21%

In those patients who were not transfused with a hematocrit < 27%, the mortality was near 50%. Naturally, those with anemia had significant comorbidities, so it could be argued that anemia was just a marker for a more ill individual who was not going to do well postinfarction. However, if this were the case, one would not expect the excellent results of transfusion. Relative risk of death in 30 days was significantly reduced by transfusion for hematocrits below 33% even when adjusted for other clinical factors (see Table 2).

Table 2

Relative Risk of Mortality by Initial Hematocrit Range in Transfused vs. Untransfused Patients

5-24% 24-27% 27-30%  30-33% 33-36% 36-39%
Unadjusted .37  .42  .49  .62   1.01 1.43
Adjusted   .31  .48  .54  .64  1.05  1.25

Interestingly, at hematocrits > 36% relative risk of death actually increased in those transfused.

The major problem with this report is that it is a retrospective, observational study in which only one quarter of patients with hematocrits < 33% got transfusion.

On the other hand, this rate falls within the range in other studies of 0.2-27%. Also, two thirds of the patients in the original database were excluded for a variety of reasons. This figure is actually low for this type of study, in which 95% often are rejected. In addition, it is not known if these results are transferable to younger patients with less comorbidities. Such patients may be able to tolerate a lower oxygen carrying capacity of the blood and could derive less benefit from transfusion.

The low rate of transfusion is undoubtedly due to the almost complete lack of data in this situation. Older studies of hospitalized patients in general suggested that adverse clinical events do not occur more frequently until the hematocrit is below 27%. Some guidelines acknowledge that patients with coronary occlusions may need more blood cells, and suggest a cut-off of 30% in the acute myocardial infarction setting. The latter expert opinion comes close to the data presented in this study. Wu et al recommend transfusion for hematocrits < 30%, but the accompanying editorial urges transfusion at 33% or less.1 Despite this study’s flaws, it is valuable to have a definite cut-off point to confidently tell our patients and colleagues when the acute myocardial infarction patient needs a blood transfusion. 


1. Goodnough LT, Bach RG. N Engl J Med. 2001;345: 1272-1344.