DURHAM, N.C. -- In a study of more than 44,000 patients being treated for a possible heart attack, cardiologists at the Duke Clinical Research Institute found that while transfusions were associated with a benefit in some patients, they were associated with harm in others.
This finding of harm with transfusions in general is not new, but extends the suggestive evidence from patients in clinical trials to "real-life" patients seen in the community, the researchers said.
The findings further suggest that providers should reconsider their decision-making process about which patients with acute coronary syndrome, a condition in which patients show signs of a heart attack, should get transfusions, said the study's lead investigator, Karen Alexander, M.D. She presented the study results on Wednesday, Nov. 15, 2006, he annual scientific sessions of the American Heart Association, in Chicago.
The researchers looked at a key blood measure, known as hematocrit, to see when physicians made the decision to transfuse blood, and they then compared this transfusion "decision point" to the health outcomes of the patients. Hematocrit is a percentage representing the amount of blood volume that is made up of oxygen-carrying red blood cells. For males, the normal hematocrit range is 42 percent to 52 percent, and for women the normal range is 36 percent to 48 percent.
Alexander and colleagues examined how hospitals nationwide treated patients with acute coronary syndrome with transfusions based on their lowest recorded hematocrit. These patients either were anemic when they arrived at the hospital or lost a significant amount of blood while being treated. In both cases, physicians typically give the patients blood transfusions.
The researchers found that while transfusions were beneficial in those whose "nadir" hematocrit, or lowest level, was less than 24 percent, transfusions were associated with greater harm in those whose "nadir" hematocrit was greater than 30 percent, Alexander said. For patients with a hematocrit between 24 percent and 30 percent, the researchers found that transfusions were associated with no benefit or no harm.
"Our data suggest that providers may want to reconsider how they decide which patients should get transfusions," Alexander said. "Our data confirms no harm or benefit in the medium range of 24 percent to 30 percent, so in this group of patients, it might be best to wait and see if the hematocrit drops farther before making the decision to transfuse. Given the scarcity of the blood supply, we certainly want to apply this therapy in those who stand to benefit the most while at the same time avoiding harm."
In the study, 1,293 patients with a hematocrit of less than 24 percent received a transfusion, while 319 received a transfusion with a hematocrit greater than 30 percent. In the middle group, 2,998 received transfusions.
For the analysis, Alexander and colleagues drew on a national database called CRUSADE, which is coordinated by the Duke Clinical Research Institute and contains patient information from more than 400 hospitals. The study was supported as part of CRUSADE, with additional funding from the National Institute on Aging.
In total, the researchers identified 44,242 patients treated for acute coronary syndrome from 2004 to 2005. Of this population, 10.4 percent had received a blood transfusion, and 3.9 percent of the patients had died, Alexander said.
In transfused patients with a nadir hematocrit of less than 24 percent, the mortality rate was 12 percent, compared with 15 percent for those who did not receive a transfusion, Alexander said. The mortality rates for patients with a hematocrit of 24 percent to 30 percent were similar whether they were transfused or not.
The reasons why transfusions may cause harm are unclear, Alexander said. The red blood cells may be depleted of nitric oxide, which helps deliver oxygen from the cells to tissues but which degrades quickly in stored blood. It also is possible that transfused blood may stimulate an immune response that exacerbates already existing heart disease. A randomized study is needed to clarify the safety and benefit of transfusion, just as other therapeutic interventions for acute coronary syndromes are tested, Alexander said.
Other researchers who participated in the study were Anita Chen, Kristin Newby, Nancy Allen-LaPointe, E. Magnus Ohman, Matthew Roe, William Boden and Eric Peterson.
CRUSADE is coordinated by the Duke Clinical Research Institute. It is funded by Schering-Plough Corp, with addition funding from the Bristol-Meyers Squibb/Sanofi Pharmaceuticals Partnership, and Millennium Pharmaceuticals. Alexander has no financial interest in the sponsors of CRUSADE.