Published in the July 2008 issue of Today’s Hospitalist
What hemoglobin level should trigger a red blood cell transfusion? And does that level vary depending on underlying disease? These were just two of the questions posed by Jeffrey L. Carson, MD, chief of the general internal medicine division at the University of Medicine and Dentistry of New Jersey-Robert Wood Johnson Medical School in New Brunswick, during a presentation at this spring’s Society of Hospital Medicine annual meeting.
According to the American Red Cross, more than 14million units of blood are transfused in the U.S. every year. But despite the common use of blood transfusions, transfusion medicine continues to be dogged by controversies and a lack of conclusive evidence. That leaves hospitalists wondering when and to whom to give blood, Dr. Carson said.
Here’s a look at some of the trickiest transfusion topics, from the use of platelets and fresh frozen plasma to the safety of erythropoietin (EPO).
Who benefits from transfusion?
To discuss the evidence behind red blood cell transfusions, Dr. Carson presented the case of an 80-yearold man who is seen two days after undergoing surgery for a hip fracture. The patient has diabetes and had a myocardial infarction three years ago. Postoperatively, he is weak, with a hemoglobin level of 8.5 g/dL. Should he be transfused?
Dr. Carson said that animal data suggest “that patients with cardiovascular disease may be less tolerant of anemia than patients without cardiovascular disease.” That association was also borne out in a study that Dr. Carson authored in the Oct. 19, 1996, issue of The Lancet. The study looked at nearly 2,000 patients with cardiovascular disease who underwent surgery and declined blood transfusion for religious reasons. While their overall 30-day mortality was 3.2%, the mortality rate in patients with preoperative hemoglobin levels of 12 g/dL or higher was 1.3%. In patients whose preoperative hemoglobin was less than 6 g/dL, by comparison, the mortality rate was 33.3%.
“As the hemoglobin falls,” said Dr. Carson, discussing the study results, “patients with cardiovascular disease have much higher odds of death compared to patients without cardiovascular disease.”
Liberal vs. restrictive transfusion strategies
Would a transfusion of red blood cells modify that risk? One meta-analysis cited by Dr. Carson published in the Cochrane Database of Systematic Reviews found that overall, blood transfusion had no impact on clinical outcomes. But the Transfusion Requirement in Critical Care (TRICC) trial, published in the Feb. 11, 1999, New England Journal of Medicine (NEJM), did show an impact.
This randomized, controlled study “which Dr. Carson said is “the only adequately powered clinical trial in the world’s literature today” “compared a “liberal” transfusion strategy (transfusion trigger of 10 g/dL, target 10-12 g/dL) to a “restrictive” strategy (transfusion trigger of 7 g/dL, target 7-9 g/dL) in critically ill patients.
Patients in the restrictive group had fewer MIs (0.7% vs. 2.9%) and less pulmonary edema (5.3% vs. 10.7%). Overall 30-day mortality was similar in both groups (18.7% for the restrictive arm vs. 23.3% in liberal arm).
“The restrictive strategy was found to be at least as effective as the liberal strategy with respect to mortality,” noted Dr. Carson. “These data suggest that a trigger of 7 g/dL is as safe as a 10 g/dL threshold in ICU patients.”
The study authors noted, however, that 30-day mortality rates among patients in the restrictive group were “significantly lower” when patients were less acutely ill. In addition, a sub-analysis of patients with ischemic heart disease found “non-significant lower mortality in patients transfused at 7 g/dL “20.5% mortality vs. 22.9% in the liberal group,” Dr. Carson said. “This raises the question of whether the threshold for transfusion should be higher in patients with cardiovascular disease.”
No answers in observational studies
Observational studies shed little light on what transfusion trigger physicians should use when treating patients with cardiovascular disease.
One study that Dr. Carson helped author in the Jan. 21, 1998, Journal of the American Medical Association (JAMA), for example, found that blood transfusions had no impact on patient outcomes, whether or not patients had cardiovascular disease.
A second observational study, which looked at MI patients and was published in the Oct. 25, 2001, NEJM, found that blood transfusion reduced mortality when the hematocrit level fell below 33%, Dr. Carson said. But it was the sole study to reach that conclusion.
And a third study published in the Oct. 6, 2004, JAMA, found that blood transfusion actually increased 30-day mortality among patients with acute coronary syndrome three- or four-fold.
“The point is that you cannot trust observational data,” Dr. Carson said. With that in mind, how would he treat the case he presented? “The best data suggest that a restrictive transfusion trigger of 7 g/dL should be used for patients without cardiovascular disease,” he noted, adding that he wasn’t sure if he would transfuse the 80-year old hip fracture patient with a previous MI and a hemoglobin of 8.5 g/dL.
Despite the contradictory evidence, Dr. Carson did offer one piece of advice to hospitalists: Ignore the old adage to give blood whenever the hemoglobin falls below 10 g/dL.
Platelets and fresh frozen plasma
Scant evidence is also a problem when deciding what thresholds to use to transfuse platelets or fresh frozen plasma.
What we know from evidence on triggers for platelet transfusion is “almost nothing,” said Dr. Carson. There are no clinical trial data testing a 10,000 vs. a 20,000 threshold for platelet transfusion, for example, except in chemotherapy.
He said that “the standard textbook stuff that we’ve been taught” is not based on randomized evidence. Current teachings hold that stable patients should not get platelets unless their platelet count falls below 10,000.
Patients with a 20,000 count are considered for transfusion if they are febrile, septic or have other bleeding issues. “The surgery threshold seems to be 50,000,” he continued, “but if you’re going anywhere near the central nervous system, people want much higher counts, up to 100,000.”
Likewise, there is no “good evidence” for establishing a certain INR threshold for fresh frozen plasma transfusions. “The textbooks basically say that if the INR is less than 1.5, you do not need to give fresh frozen plasma,” Dr. Carson pointed out. Hemostasis can be achieved when the activity of coagulation factors is at least 25% or 30% and fibrinogen is at least 75 to 100 mg/dL.
“The key mistake commonly made with fresh frozen plasma is that doctors don’t give as much as they need to,” Dr. Carson said. “You need to administer more than you would normally give, between three and five units, for example, and you need to watch for volume overload.”
EPO: Is it safe?
Dr. Carson also discussed the risks of another controversial therapy for anemia: EPO, or erythropoiesis stimulating agent. Where does this therapy fit into hospitalists’ armamentarium, and what issues should doctors worry about?
The FDA has approved erythropoiesis stimulating agents to treat anemia caused by chronic renal insufficiency, cancer, HIV and surgery. Although the medications are commonly used in the ICU, they are not approved for use in ICU patients.
Questions about the safety of these agents were raised after the FDA recently issued new black box warnings noting that the drugs “increased the risk for death and for serious cardiovascular events when dosed to achieve a target of hemoglobin of greater than 12 g/dL.” Those warnings were based on a meta-analysis of nine randomized controlled trials of patients with chronic renal insufficiency, published in the Feb. 3, 2007, issue of The Lancet.
According to Dr. Carson, the study identified a “17% increased risk of death in patients treated with ESAs, and the risk was primarily identified in studies where hemoglobin was normalized above 12 g/dL.” In addition,” he said, “the risk of developing a blood clot was 34% greater in the higher hemoglobin group.”
“The current evidence suggests that the risk of maintaining hemoglobins above 12 g/dL outweighs the potential benefit,” he added. A study of ICU patients published in the Sept. 6, 2007, NEJM found that the use of epoetin alfa did not reduce the incidence of red blood cell transfusion or improve overall 30-day mortality.
Dr. Carson noted, however, that there was a statistically significant mortality reduction in a subgroup of trauma patients. The only significant adverse effect was an increased incidence of deep vein thrombosis.
“The data are really pretty consistent in showing that these drugs are associated with a higher risk of clot, including in the ICU setting,” Dr. Carson said. As for efficacy data, those show that “perhaps there is a role in trauma, but that needs to be replicated.”
Perhaps more disturbing, noted Dr. Carson, is evidence that EPO may promote cancer recurrence. A study published in the Oct. 18, 2003, issue of The Lancet looked at cancer progression and recurrence in patients with advanced squamous cell carcinoma of the oropharynx, who had initial hemoglobin levels less than 12 g/dL.
EPO was given to reach a target hemoglobin of 14 g/dL for women and 15 g/dL in men. While it was a relatively small study, investigators found “a 62% increased risk of recurrence,” Dr. Carson said. “Survival was also adversely affected.”
While indications for EPO are still emerging, he concluded, “at this point in the setting of cancer, you should be careful using it.”
Thumbs down for autologous blood
Finally, hospitalists at the presentation raised the issue of whether or not to use autologous blood for transfusions. According to Dr. Carson, autologous transfusions should be eliminated from clinical practice.
“When you give autologous blood, you have many of the same concerns as with other units of blood, and you end up wasting 50% of your units,” he pointed out. If you do use autologous blood, he added, “use the same restrictive transfusion thresholds as with donated blood.”
Cornelia Kean is a freelance health care writer based in Montclair, N.J.
A look at transfusion risks
A 65-YEAR-OLD with a history of coronary artery disease and chronic obstructive pulmonary disease is admitted with gastrointestinal bleeding, a hemoglobin level of 7 g/dL and an INR of 1.5. She is given two units of blood and three units of fresh frozen plasma, but two hours later is dyspneic with diffuse alveolar infiltrates shown on chest X-ray. What is the diagnosis?
According to Jeffrey L. Carson, MD, the patient is suffering from transfusion-related acute lung injury (TRALI), one of the newly recognized risks of transfusions. Although it has only recently been recognized, experts estimate that TRALI occurs in one in 5,000 red blood cell transfusions.
Dr. Carson, chief of general internal medicine at the Robert Wood Johnson Medical School in New Brunswick, N.J., says typical symptoms include chills, fever, a non-productive cough, dyspnea, cyanosis, and hypotension or hypertension occurring within one or two hours of transfusion.
“These patients look like they have an ARDS-like picture,” he explained during a presentation at the Society of Hospital Medicine annual meeting. “They are not in heart failure, which is key to your differential diagnosis.”
While the cause of the reaction is not completely understood, experts speculate that it may be linked to antileukocyte antibodies and bioactive lipid factors present in donor products. According to the National Institutes of Health, these proteins “which are more likely to be found in the plasma of women who have been pregnant “may disrupt normal lung-cell function.
Because of this risk, hospitals in the United Kingdom “have removed female donors from their fresh frozen plasma pool,” Dr. Carson reported. “They believe the rate of TRALI has fallen substantially.”