September 2012
Feature Story

Anne Paxton

When the TRICC trial (Transfusion Requirements in Critical Care) in 1999 brought the first evidence from a randomized, controlled study that critical care patients can do just as well with a restrictive strategy of transfusion as with a liberal strategy—and in some cases, they do better—it was a watershed of sorts for transfusion medicine.

But it didn’t exactly mark a turning point in clinical practice. A series of important studies later confirmed similar findings, and a few specialty societies have issued general recommendations for lower transfusion thresholds. Still, between 1997 and 2009, the number of all hospital stays with a blood transfusion included jumped from 40 per 10,000 to 92 per 10,000—a cumulative growth rate of 128 percent—while supplies, at 15-year lows earlier this summer according to the American Red Cross, struggle to keep pace. Aside from labor and delivery, blood transfusion remains the most common procedure that an adult receives in the United States.

A few months ago, the AABB, formerly the American Association of Blood Banks, took a step that many are hoping will tip the scales in favor of restrictive transfusion strategies. In a March 26 article in the Annals of Internal Medicine, the group issued the first clinical practice guideline on RBC transfusion representing the consensus of the transfusion medicine community. Significantly, it is also the first such guideline to recommend a specific transfusion trigger for hospitalized, stable patients: a hemoglobin of 7 g/dL to 8 g/dL. Could the AABB guideline finally provide that last nudge needed to encourage most clinicians to ratchet down their blood use?

Based on a systematic review of the literature on randomized clinical trials evaluating transfusion thresholds, the AABB guideline uses evidence-based procedures to make essentially three recommendations. First, the guideline recommends adhering to a restrictive transfusion strategy (7 to 8 g/dL) in hospitalized, stable patients. Second, the guideline suggests adhering to a restrictive strategy in hospitalized patients with preexisting cardiovascular disease and considering transfusion for patients with symptoms or with a hemoglobin level of 8 g/dL or less. Third, the guideline suggests that transfusion decisions be influenced by symptoms as well as hemoglobin concentration.

Noting that further research is necessary, the AABB declined to recommend for or against a liberal or restrictive transfusion threshold for patients with acute coronary syndrome.

The three recommendations in the guideline are the product of the AABB’s Clinical Transfusion Medicine Committee, which has had groups of physicians from clinical settings, including hematologists and pathologists, working on guidelines for the last five years. The RBC guideline is the second of three planned by AABB; a plasma transfusion guideline came out in 2010, and development of a platelet guideline is just getting underway.

“AABB wanted to take a leading role in guideline development, and our board identified this activity as a strategic priority a number of years ago,” says Theresa Wiegmann, JD, AABB director of public policy. Part of the guideline development process was ensuring the involvement of specialists in critical care medicine, trauma surgery, and other areas. “We tried to pull in other specialties that frequently transfuse blood products, such as anesthesiologists and, in the case of the RBC guideline, cardiologists.”

The AABB believes there is a great need for more consistent transfusion practices, Wiegmann says. “There are wide discrepancies in the use of red blood cells. For example, a study a few years back on the prevalence of RBC transfusions in coronary artery bypass graft surgery procedures [JAMA. 2010;304:1568–1574] looked at how many of them in a particular hospital would involve a transfusion, and it ranged from under 10 percent to over 90 percent. These discrepancies in how physicians and hospitals are making decisions on when to transfuse highlight the importance of providing better guidance about when it’s appropriate to provide blood.”

“Several well-known guidelines have been previously published,” says Jeffrey L. Carson, MD, chief of the Division of General Internal Medicine, UMDNJ-Robert Wood Johnson Medical School, New Brunswick, NJ, and lead author of the AABB guideline paper in Annals. He notes that the American Society of Anesthesiology, as well as British, Australian, and New Zealand societies, have provided general recommendations, but not a specific transfusion trigger, while other recent guidelines have recommended avoiding transfusion based only on a hemoglobin trigger. But because the AABB committee had the advantage of a larger body of randomized clinical trial data, and limited its analysis to that, it had an advantage. “Our orientation was different and our guidelines are a bit more specific. We basically said, ‘Here is the best evidence in the field and we’re going to develop a set of guidelines only based on this high-quality evidence.’”

In addition to chairing the AABB guideline effort, Dr. Carson led the 2011 FOCUS trial on transfusion of hip fracture patients (Liberal or restrictive transfusion in high-risk patients after hip surgery. NEJM. 2011;364:2453–2462), which turned out to play a central role in the guideline.

“We enrolled patients after surgery when the hemoglobin level was less than 10 g/dL. These patients were hemodynamically stable, in contrast with the trauma patient who is hit by a car and comes in hemorrhaging. Some patients had anemia from other causes, but we did not collect this information. However, we are sure that the causes of anemia in the two transfusion groups will be similar because the trial included 2,000 patients. We do not believe this is an important issue,” Dr. Carson says.

In an earlier systematic review, “There had been about 3,700 patients who had been randomized in mostly small clinical trials. We updated the review after the FOCUS trial, and one other trial was published last year. This increased the number of patients to about 6,200. So it was an ideal time to look at the evidence again and see if we could develop a guideline.”

The AABB group decided against recommending a broad range of transfusion thresholds, such as hemoglobin levels of 6 to 10 g/dL. “We recommend for ICU patients using a threshold of 7 g/dL because the TRICC trial showed this, and there’s a pediatric equivalent trial [TRIPICU] that showed 7 as well. In patients who are likely to be comparable to the hip fracture patients in the FOCUS trial, we recommend ‘8 g/dL or symptoms.’ We don’t recommend 7 because the trials did not test this threshold. We used the evidence from the trial, showing that 8 g/dL or symptoms seems to be as good as a threshold of 10 g/dL, so therefore we think you can consider transfusing that patient. However, we advised to take into account the patient’s symptoms, not just the hemoglobin level.”

The guideline is focused on the clinical trial evidence in the populations studied. “We do extend those results to other populations where we think it’s reasonable to do so, but we were pretty careful in the way we worded things to keep it narrow,” Dr. Carson says.

There is a shortage of evidence for the group for which the AABB guideline decided not to make a transfusion threshold recommendation: patients with acute coronary syndrome. But Dr. Carson has led a clinical trial, called MINT (Myocardial Ischemia and Transfusion), a 110-patient pilot study funded by the National Institutes of Health, which looked at transfusion thresholds in patients with acute coronary syndrome. A manuscript about that study is now being readied for submission for publication. “That’s a group of patients where our guidelines say ‘we don’t know what to do.’ They may benefit from more blood, but it hasn’t been studied.” Researchers are also awaiting results of a large trial being done in the United Kingdom on cardiac surgery and transfusion. “That could change what we conclude. We’ll have to see what it shows,” Dr. Carson says.

The decision to be specific about triggers was carefully considered by the AABB Clinical Transfusion Medicine Committee. “We have all grown up saying not to transfuse based on a trigger but on the patient’s symptoms. But we thought it was important to have guidelines addressing a trigger—because that’s the way some physicians transfuse,” says guideline committee member and co-author Brenda J. Grossman, MD, MPH, medical director of transfusion medicine services, Washington University School of Medicine. “We don’t want people to use the guidelines blindly and ignore symptoms, but we wanted to consider practices.”

That approach has generated debate. In a related editorial published alongside the AABB guideline, Jean-Louis Vincent, MD, PhD, Department of Intensive Care, Erasme Hospital, Free University of Brussels, Belgium, warns that assigning specific triggers can be risky. In his view, the decision to transfuse is too complex and important to be based on single numbers.

Dr. Vincent does not challenge the conclusions reached by the AABB guideline developers. He writes: “It would be difficult to argue with their statement that, ‘On the basis of data from all the randomized trials, the panel found little evidence to support a liberal transfusion strategy.’ Certainly, ample evidence supports the assertion that, in general, less may be better when it comes to blood transfusion.” However, he does not believe that available evidence supports a fixed transfusion trigger. “Rather, transfusion decisions need to consider individual patient characteristics, including age and the presence of coronary artery disease, to estimate a specific patient’s likelihood of benefit from transfusion.”

In Europe, the transition from a general policy of liberal transfusion to the now current more restrictive strategies has been a gradual one, said Dr. Vincent in an interview with CAP TODAY. “Over time, we’ve learned that the rule which was in favor several decades ago to keep everybody above 10 g/dL was too simplistic. So progressively over the years, we started to think about benefits but also the risks of transfusions in every single patient we need to evaluate.”

Most physicians genuinely try to take into account multiple factors before transfusing, says Jerome L. Gottschall, MD, senior medical director of the BloodCenter of Wisconsin and chair of the CAP Transfusion Medicine Resource Committee.

“My own opinion is that physicians do their best to transfuse RBCs based on the hemoglobin level, the hematocrit, and the clinical status of their patients,” but the guideline has tried to point out that you can keep patients at a lower hemoglobin level than probably many physicians would realize, he says. He thinks this guideline will be well received by physicians.

Asignificant feature of the AABB guideline development process is that it was strictly evidence-based, using the GRADE method, and referred only to rigorous clinical trials, not observational trials. “These guidelines make a wonderful attempt to use evidence-based medicine and to use clinical trials that were somewhat randomized, and they pointed out particularly that they did not try to use any observational studies that have the potential for more bias in them,” Dr. Gottschall says.

For recommendation No. 1 (7 g/dL threshold for stable hospitalized patients), the authors note, there was high-quality evidence. “We decided that if everything were equal—in other words, there was no evidence it was harmful to restrict and the evidence all indicated a restrictive strategy was better or the same—we would choose a restrictive strategy,” Dr. Grossman says. In fact, for recommendation No. 1 there was no evidence to the contrary. “Although transfusion triggers differed among trials, no results favored the liberal strategy,” the authors note. “In fact, the three largest trials conclusively showed a lack of benefit with liberal transfusion.”

For recommendation No. 2 (8 g/dL threshold for hospitalized patients with preexisting cardiovascular disease), the evidence was ranked moderate quality, in part because the clinical trial data were limited. Recommendation No. 2 was backed by a couple of trials including the 2011 FOCUS trial, which included patients with cardiovascular disease and cardiovascular risk factors. “Overall,” the guideline authors note, “there was no difference in functional recovery, mortality, or hospital complications of myocardial infarction, congestive heart failure, stroke, infection, or thromboembolism with a liberal or restrictive transfusion strategy.” However, Dr. Grossman notes, in the FOCUS study the restrictive group was being transfused with a hemoglobin of 8 g/dL or less, or symptoms, “so it was not possible to make a recommendation with a 7 g/dL threshold in the guideline.”

It was similar concern for the evidence that led the AABB to decide against a recommendation on thresholds for hemodynamically stable patients with acute coronary syndrome. While observational studies in patients with ACS have been reported, “because of uncontrolled confounding, the panel believed that the evidence from these types of studies was insufficient to support clear recommendations,” the guideline says.

The last recommendation, to make transfusion decisions be influenced by symptoms as well as hemoglobin concentration, actually relied on the FOCUS trial, which was the only trial that incorporated symptoms into the decision to transfuse. Overall, in that trial, transfusions for symptoms were reported in 15.7 percent of patients in the restrictive group versus 5.3 percent in the liberal group, but there were not statistically significant differences in any of the outcomes examined.

The AABB panel noted that an optimal trial would compare patients receiving transfusions based on hemoglobin concentration with those receiving transfusions only for symptoms. But at the same time, the panel said, it believed it unlikely that such a trial will be done: “Although such a trial would be valuable, it would probably require blinding physicians to the hemoglobin concentration, which raises issues of feasibility. Furthermore, there is probably a lack of clinical equipoise about transfusion in symptomatic patients.”

Despite such reservations expressed in the guideline paper, Dr. Grossman believes physicians are much more used to a greater focus on quality of care in today’s clinical environment. “I think everybody will be a whole lot more receptive nowadays than in the past,” she says. “It’s very difficult when you look at the systematic evidence-based method used to come up with the guideline, to dismiss it now, where in the past when quality was not as visible, they may have ignored it.”

The guidelines will be useful to hospitals and other institutions, Dr. Gottschall agrees. “People have been working across the country to improve the practice of transfusion therapy for many years, and the development of good guidelines from important organizations is really helpful in establishing guidelines in one’s hospital.”

As a preview of the potential impact of hospitals’ adopting the AABB guideline recommendations, the experience of the University of Alabama at Birmingham could be instructive. In 2006, demand for RBCs was definitely exceeding supply, says Marisa B. Marques, MD, professor in the Department of Pathology and coagulation service director, and a member of the AABB Clinical Transfusion Medicine Committee who worked on the guideline. “We had reached a demand for RBCs of 40,000 units a year, which is very difficult to obtain. Blood centers had to scramble to provide us with our standing orders of so many units. For many years before that, we had seen an increase in usage every year between five and seven percent.”

Dr. Marques had tried to work with the physicians in the specialties that used blood the most to see if anything could be done to avoid shortages. “By myself, I had not been able to bring this about; everybody thought we were using units as reasonably as we could.” The hospital brought in a consultant who determined the institution’s triggers for transfusion were much higher than the best evidence in the literature.

Based on the TRICC study and subsequent studies, the hospital adopted a recommended Hb level of 7 g/dL as a transfusion trigger for acutely ill patients except those with rapid blood loss, with acute coronary syndrome, or on sepsis protocol. The program includes ongoing retrospective review of units transfused outside the guideline, and the results have been impressive: a steady drop in the use of RBCs from 40,043 units in 2007 to 29,877 in 2011, a total 25 percent decrease over four years. “But I know we still have room to improve further,” Dr. Marques says.

The same evidence was used to create the AABB guideline, but now there is even more of it, she notes. “The more transfusions of hospitalized critically ill patients we see, the more morbidity and mortality they have. But physicians have not necessarily realized there are severe risks related to transfusion beyond HIV and HCV.” The other risk of unneeded transfusion is that blood won’t be available when needed. “The more units we use inappropriately, the fewer units we have to use in someone who truly needs them.”

In the whole country, the impact of a similar program is hard to estimate because each hospital is at a different stage, Dr. Marques says. “Some may be able to reduce 10 percent, some 30 percent, others 60 percent. But most definitely, I can guarantee, we can decrease overall need.” Pathologists must lead the way, she says. “Every pathologist in charge of a transfusion service in their own hospital should be updated with the guidelines and help their clinical colleagues understand the pros and cons of transfusion. It’s our job to disseminate the information, because our clinical colleagues have a lot of different aspects of medical care to keep up with.”

Other hospitals could take the same road as Dr. Marques’ hospital, says Wiegmann of the AABB. “There is a bit of a trend toward implementing patient blood management in hospitals, and we are hopeful that facilities will follow the guideline. This is an integral part of patient blood management, to transfuse only when necessary, not only improving patient care but also helping hospitals limit costs.”

At least some speculation has surfaced that hospitals following AABB recommendations could reduce RBC transfusions by as much as 30 percent to 40 percent. “That would be pretty amazing,” Dr. Carson says. “Most guidelines don’t have that kind of impact. But here’s what I can tell you. If you look at the two different groups in the FOCUS trial, the group that was randomized to receive less blood got one-third the amount of blood of the other group. That doesn’t mean we’re going to reduce blood use by that much, because it implies all doctors in the U.S. are using a lot of blood. And that’s not true.”

Many additional factors are in play, Dr. Carson notes. “The patients who are admitted to our hospitals are older and sicker, and they’re undergoing more complicated procedures, so I don’t know.” Still, he has seen indications that the transfusion landscape might already be changing. “A medical director of a large medical group tells me every year they have planned for an increased amount of RBCs, and this is the first year there’s been a reduction of about three percent. That’s not much—but it’s not an increase.”

There are a number of areas where the answer is not yet in, Dr. Gottschall says. “Patients with acute myocardial infarction, patients having gastrointestinal bleeding, and patients with brain injury are three examples. And there are not enough studies in truly elderly patients who are hospitalized. Exactly where should they be? And what is the best level to help them go home and be comfortable? These are areas where we simply don’t have evidence encompassing a lot of patients, and that makes it difficult to decide where to keep a hemoglobin level at.” But he is hopeful that the recommendations of AABB’s guideline will affect the way transfusion strategies are deployed in hospitals and taught in medical school and residency programs. “I can tell you I’ve already taken these guidelines and incorporated them into the talks that I give.”

Dr. Carson is cautiously optimistic. “What happens with physicians, and I think any rational people, is they’re used to doing things one way,” he says. “Then they get a study that says maybe this isn’t right. And then when that study is replicated in a high-risk group of patients, the evidence becomes much more convincing and compelling and things start to change. We’re hoping that the guidelines will pile onto that, and we’ll see if they change practice patterns.”