October 2009
Feature Story

Anne Paxton

Faced with unrelenting pressure to stamp out threats to the safety of the blood supply, the nation’s blood centers can’t afford to coast on their record or let down their guard. Nevertheless, they might just be breathing a tiny sigh of relief at their recent progress in reducing cases of TRALI (transfusion-related acute lung injury), the leading cause of death from transfusion.

In 2006, the AABB recommended that its members—which include most blood centers in the U.S.—adopt measures to address TRALI through plasma transfusion by November 2007, and TRALI by apheresis platelets by November 2008. The specific steps to take were left up to individual blood centers. But large numbers of them opted to move to predominantly male plasma. “In essence, using predominantly male plasma for transfusion is now the industry practice; it—or its equivalent—has become the standard of care,” says James AuBuchon, MD, president and CEO of Puget Sound Blood Center, Seattle.

The results have been dramatic: steep and relatively rapid declines in deaths from TRALI. “Over the five years from 2003 to 2007, TRALI was the No. 1 cause of transfusion-related death in the U.S. reported to the FDA. In 2008, interventions to reduce the risk may have had an impact, resulting in fewer cases reported—16 in 2008 versus 34 in 2007,” says Darrell Triulzi, MD. This “is now about the same as the number of deaths reported from hemolytic transfusion reactions.” Dr. Triulzi is medical director of the Institute for Transfusion Medicine, Pittsburgh.

“TRALI remains the leading severe and fatal risk of transfusion,” emphasizes Richard Benjamin, MD, PhD, chief medical officer of the American Red Cross, which has moved its 36 blood regions to predominantly male plasma. “But we have now very strong data that moving to predominantly male plasma for transfusion has been an effective measure in decreasing TRALI.” At the Red Cross, “we have seen an approximately 70 percent reduction in reported cases of severe TRALI.”

Indeed, the Red Cross is reporting at this month’s AABB meeting in New Orleans that in 2008, it received zero reports of fatal cases of TRALI caused by plasma without any other component, where previously it saw six to eight cases a year. But U.S. blood centers still face challenges in their attempts to reduce TRALI from platelets and red cells, and there remains much about TRALI and its pathogenesis that medical science has not uncovered.

TRALI, a syndrome defined as a new acute lung injury within six hours of a completed transfusion, is a rare event and the average clinician is likely to encounter one case of it in his or her whole career, Dr. AuBuchon says. “But when you see a patient with TRALI, you’ll remember it. Their oxygen saturation drops, fluid accumulates in their lungs where air is supposed to be, and they’re gasping for breath. It presents a very dramatic clinical situation.” About 10 percent of the time, it leads to death.

It was the United Kingdom that first decided to go with an all-male plasma policy, beginning in 2004, and subsequently witnessed a sharp downturn in TRALI cases. (What has resulted so far has been predominantly male owing to operational considerations. The U.K. is now at 98 percent male plasma.) “The U.K. did an analysis through its hemovigilance program where they looked at data from the late 1990s to early 2002 and observed that TRALI events were associated with transfusion of plasma blood products from women, so they went ahead and decided to intervene and make plasma from exclusively male donors,” says Robert Makar, MD, PhD, assistant director of the blood transfusion service at Massachusetts General Hospital and an instructor at Harvard Medical School.

That intervention brought swift results. There were no cases of TRALI attributed to fresh-frozen plasma in the years 2005 through 2007, according to Lorna Williamson, MD, medical and research director of National Health Service Blood and Transplant in England. “Unfortunately, there were three cases in 2008 due to having to make some female fresh-frozen plasma to maintain stocks,” she says, and adds, “NHSBT is committed to a consistent 100 percent male FFP policy from mid-2009.”

Patricia Kopko, MD, chief medical officer and histocompatibility director at BloodSource in Sacramento, and assistant clinical professor of medical pathology at the University of California at Davis, says these results demonstrate that blood centers were right to take action early—even before all the data were in.

It might not have happened if evidence-based medicine concepts were strictly followed, she points out. At a recent meeting she attended, “Many medical experts stated that you shouldn’t do risk reduction measures without evidence-based medicine. But how do you do evidence-based medicine for rare events? You don’t. You can’t.” Especially with the incidence of TRALI now down to perhaps as low as one in 10,000—“How do you do randomized, controlled studies? No institutional review board on earth would approve transfusion of plasma known to have HLA or HNA antibodies to a patient.”

The Community Blood Center of Greater Kansas switched to all-male plasma in March 2007 after seeing the SHOT (Serious Hazards of Transfusion) data in the U.K., says Jay Menitove, MD, president and CEO and medical director. “We’d like to have no cases, and it would be nice to know all the reasons for TRALI. But based on current knowledge, we’ve acted prudently, consistent with what the data are showing, and we’ve seen a decline in cases consistent with what the U.K.’s SHOT data suggest.”

Although researchers do not have a definitive understanding of the causes of TRALI, the syndrome’s likely causes are emerging from transfusion medicine research. Experimental and epidemiologic data of the past couple of years support the theory that TRALI is antibody-mediated, Dr. Triulzi says.

“There is more confidence that TRALI is primarily caused—we believe in at least 80 percent of cases—by donor leukocyte antibodies, which are either HLA or neutrophil antibodies,” he says. Dr. Triulzi is directing the multicenter leukocyte antibodies prevalence study (LAPS), sponsored by the National Institutes of Health, to examine the frequency of such antibodies in 8,000 blood donors.

Many of the other cases might be explained by an alternative model for how TRALI happens: the bioactive lipid model. “This model proposes that cellular blood products, with time, accumulate breakdown products of membrane lipids that in the right clinical context can cause acute pulmonary injury,” Dr. Makar says. “However, the lipids themselves are not sufficient to cause TRALI; clinical factors such as inflammation are also needed to provide another ‘hit’ that puts the patient at risk for this kind of response to transfusion.”

As possible overall explanations, variations of this “two-hit” model are gaining adherents. In the first hit, the patient’s white cells, neutrophils, are primed by some under­lying condition, such as sepsis, surgery, or massive transfusion, explains Mark Popovsky, MD, vice president and chief medical officer of Haemonetics, Braintree, Mass., and associate professor of pathology at Harvard Medical School, who, with Breanndan Moore, MD, first described TRALI and its features in the early 1980s.

“By priming, they basically adhere to the underlying epithelial cells in the lung microvasculature. The second hit causes the white cells to release their biochemical contents, damaging the underlying endothelial cells and causing a protein-rich fluid to leak into the lungs and cause oxygen desaturation,” Dr. Popovsky says.

However, there is increasing recognition that both anti-leukocyte -antibodies and bioactive lipids may function as a second hit in the two-hit model. “Over the last two years there has been some cross-pollination between the antibody and the bioactive lipid model,” Dr. Makar says. “Even in the antibody model, there may be clinical conditions that predispose the patient to TRALI.”

Recent studies in a rat model showed that anti-leukocyte antibodies did not cause TRALI unless the animals were given lipopolysaccharide to cause proinflammatory changes in the rat. It’s not entirely clear what clinical conditions would predispose a rat or a person to get TRALI, he adds, “but it’s very clear that just being infused with blood product that contains HLA antibodies—even when you have the cognate HLA antigen—doesn’t always mean you’ll get TRALI. Now we’re just trying to get our hands around what clinical factors seem to play a role.”

The September 2009 issue of the AABB’s journal Transfusion (Triulzi DJ, et al. 49:1825–1835) contains a report of the initial findings of the leukocyte antibodies prevalence study: Pregnancy is a strong immunizing event for HLA antibodies, with each successive pregnancy increasing their prevalence. But the data show that a history of transfusion is not a strong risk factor for HLA antibodies, Dr. Triulzi says.

There are probably several reasons for this, he believes. “Transfusion of blood donors is almost always a distant event occurring on average more than 10 years in the past. So if there ever were antibodies, it is possible that they’re gone now. We compared HLA antibodies in approximately 1,000 men who had been transfused and 1,000 who had not, and we found a minimal difference [1.7 percent in men with a transfusion; one percent in men with none]. Women who had never been pregnant or transfused also had a 1.7 percent rate.”

Such findings have direct implications for donor screening. “These data can allow blood centers to model what the impact of HLA testing would be,” Dr. Triulzi explains. “They realize it’s not necessary to ask the donor questions about transfusion; you only need to ask questions about pregnancy.”

Epidemiologic data have strongly suggested that diverting plasma units from multiparous women from the blood supply has resulted in a significant decline in TRALI cases, Dr. Triulzi says. But even one pregnancy creates an 11 percent chance of having antibodies, so blood centers are considering or are doing HLA antibody testing on women who report any pregnancy, and they are diverting plasma from women with multiple pregnancies to fractionation instead of direct transfusion.

“The technology to test for HLA antibodies has been around in high-throughput settings for several years, either by ELISA or by Luminex flow cytometry, so that was feasible, and we chose the Luminex platform for the LAPS study,” Dr. Triulzi says. With HNA antibodies—the other type implicated in TRALI—the situation is different. “We found a prevalence of about one percent and no up-front demographic factor that you could use to predict who might have HNA antibodies. Since the test is manual and laborintensive, our data would not suggest, at this point, that screening for HNA would be practical,” he says.

The followup to the leukocyte antibodies prevalence study, called LAPS II, is following plasma and platelets from HLA-positive donors to recipients to see if they developed TRALI. The data collection is scheduled to be complete in December, and Dr. Triulzi expects results will be reported in 2010.

Researchers at the University of California at San Francisco, funded by the National Heart, Lung and Blood Institute, are studying TRALI in mice and humans. Pearl Toy, MD, director of the Transfusion Medicine Specialized Center of Clinically Oriented Research and professor of laboratory medicine, is directing the project. “By having basic and clinical investigators working together, we can translate ideas generated from mouse studies quickly to human studies,” Dr. Toy says.

In this study, in conjunction with Mayo Clinic, “we will determine TRALI incidence as well as risk factors in recipients and in donor blood,” she says. Patients who developed TRALI will be compared with controls who were transfused a similar number of units of blood but did not get TRALI.

Blood centers have generally settled on two main ways to meet AABB recommendations, Dr. Popovsky says. “The first is to take female plasma out of fresh-frozen plasma production. The second is to screen female donors for HLA antibodies and if they are positive, not use their blood for plasma production.”

Female plasma may still go for fractionation—where it is pooled in vats of 10,000 liters or more, then used to create IVIG, factor VIII albumin, or antithrombin III, but not transfused to individual donors. “The process of transmission is different with fractionation,” Dr. Popovsky explains, “probably because the antibody is diluted out enough in large quantities.” In the literature, he adds, a few cases of TRALI have been associated with fractionated products, but the risk is minuscule.

A third route that some centers have taken is to use donors’ medical history as a guide—by screening the antibodies of women who have had one or more pregnancies. “There are a lot of approaches, depending on economics or ease of implementation, but the key point is that the U.S. has migrated now to a policy where male plasma and female plasma are being treated in a different way, and we’re seeing results now,” Dr. Popovsky says. It’s a community practice even if there isn’t an official standard.

The persistent need for AB plasma, however, means blood centers face possible shortages. “We do have supply issues with some types of plasma, such as group AB,” Dr. Benjamin of American Red Cross notes. “Only about three percent of the population is group AB because that’s a universal plasma type, but nine percent of the plasma we distribute is AB plasma.”

Like many other blood centers, Dr. Triulzi’s uses AB plasma from women of childbearing age to meet the need, but if it has to use plasma from women with a previous pregnancy, it will first use plasma from those with only one pregnancy.

“You can’t get enough AB plasma,” Dr. Kopko agrees. But in her experience, donors who are group AB know how rare their blood cells are and that makes them even more committed to donate. “We expire less than one percent of the red cells we collect but half are AB, so we convert the donors—currently just the males—to plasma donors since they are universal and their plasma can go to anyone. We contact those donors and ask them to donate plasma.”

If the entire country switched to all-male plasma, “I don’t think there would be a significant problem with shortages,” Dr. Popovsky says. “There are 3 to 4 million units of FFP transfused each year but 15 million collected.” In the U.K. too, there have been no significant plasma shortages created by TRALI prevention, Dr. Williamson confirms.

But to address the risk of TRALI in platelets, U.S. blood centers come up against a stark fact: They can’t afford to divert all women, because there wouldn’t be enough apheresis platelets to meet patients’ needs.

At Dr. Menitove’s blood center, for example, “For platelets we have a smaller donor base containing a significant number of female donors, of which we’d lose about 20 percent, so there are going to be inventory problems.”

Similarly, while the Red Cross already has 70 percent of its apheresis platelets coming from male donors, “we’re taking a slower approach to platelets, because we’re more concerned we’ll have a supply issue if we do a radical intervention,” Dr. Benjamin says. “So we’re introducing our measures over time.”

It won’t be quite as easy to follow the U.K.’s lead when dealing with apheresis platelets, Dr. Makar says. “It’s difficult for us to reproduce what they do because the platelet supply in the U.S. is primarily derived from single-donor apheresis products. We don’t use pooled whole-blood platelet concentrates nearly as much as they do in the U.K.”

The challenge platelets present is one reason why the AABB has, since making its initial recommendations for action on TRALI, “issued a clarification that both emphasizes its commitment to TRALI risk reduction while recognizing that implementation of measures for products such as apheresis platelets will be difficult,” Dr. Makar says.

To figure out how to manage TRALI prevention at Massachusetts General Hospital’s blood bank, Dr. Makar and colleagues did a study of blood donors who agreed to report whether they had been previously pregnant, how many times, whether they had had abortions, and what kind of blood exposures they had in the past. “We compared that data with the results of an HLA antibody screening test of their blood. Then we tried to see how that data could inform our screening strategy.”

Dr. Makar notes that the transplant world has been using a semiautomated screening test for HLA antibodies for some time and it is the kind of platform that could be used for blood screening. “Once you’ve spent capital for the actual analyzer, your reagent cost per test is about $10.”

The major conclusion of the paper, also published in Transfusion, was: “If someone presented with a history of being pregnant, you could just test only those people, and you could avoid completely removing women from the donor population and facing loss of blood products” (Powers A, et al. 2008;48:2549–2558). This study agreed with the findings of Dr. Triulzi’s leukocyte antibodies prevalence study that even a single pregnancy in a donor’s history significantly raised the risk of TRALI.

“I think in the 1980s, the data suggested that you needed multiple pregnancies to generate the antibodies that might result in a product that could induce TRALI, so the focus was on multiparous women,” Dr. Makar says. “With more sensitive testing methods using Luminex technology, we were seeing 24 to 25 percent alloimmunized after one pregnancy and something on the order of 40 percent after the second. Then the numbers tapered off.”

As a result, the blood bank decided it was cost-effective to reduce TRALI risk from plasma by simply asking if the donor has had a pregnancy or a transfusion, then screening for HLA antibodies, and not using the plasma if the test is positive.

The platelet inventory at Massachusetts General, as elsewhere, is precariously balanced, Dr. Makar says. “The platelet supply is already very, very tight because platelets are a product that only lasts five days. And there’s a tremendous need for platelet products to support patients who are thrombocytopenic—due to chemotherapy, for instance.”

“Forty percent of our platelet products are from females, and half of our females have been pregnant. Clearly we’re not going to get rid of those donors.” That would be a terrible hit, he says. Even without deferring females, the hospital has found it necessary to triage platelet products every couple of weeks. “And I’m sure others are doing the same.”

Using HLA screening as an alternative, Dr. Makar says, only about 10 percent of the hospital’s platelet product would be lost. “But even that is a little difficult for us to make up.” As a result, the blood center is thinking about “different strategies we might use to buffer ourselves when we go ahead and start deferring platelet apheresis donors. And that’s our dilemma; it’s kind of a microcosm of what’s happening everywhere.”

BloodSource has just started using the HLA antibody screening test, Dr. Kopko says. “We just started in April testing female platelet donors who have been pregnant three or more times. We know if you tested everybody at once, the positives would be so high we wouldn’t have platelets. The AABB says you have to balance the risk of TRALI with your platelet supply, so our intention is eventually to screen all female platelet donors who have ever been pregnant, but that will take some time to implement.” Because of the body size and blood volume requirements for platelet donation, only about 30 percent of BloodSource’s platelets are donated by females, she says.

Most institutions don’t yet have a big HLA lab, but BloodSource and Puget Sound Blood Center do have full-service ones. “That’s why we’re both more in the forefront of taking these measures,” Dr. Kopko says.

Of course, screening for HLA antibodies addresses only part of the picture. “Not being able to test for HNA antibodies, which are specific to granulocytes and can also cause TRALI, is not optimal,” Dr. AuBuchon says. As of yet, there’s no test amenable to the volume a blood center would need. “HNA antibody testing is expensive; it’s meant as a diagnostic test, not a screening test, and I believe doing tens or hundreds a day would just not be feasible.”

The risk of TRALI in red cells is yet another issue. It’s thought that the presence of plasma in red cells is what creates this risk. In the U.K., “We’ve been trying over the last few years to squeeze off more and more plasma from red cells to reduce the risk of vCJD [variant Creutzfeldt-Jakob disease],” which remains a leading concern for British blood centers, Dr. Williamson says. “We think most of the prions in blood are either in white cells or plasma. But we do see an occasional case of antibody-positive TRALI with red cells—maybe one case every year or two years, and those numbers aren’t really changing.”

Data about the incidence of TRALI, however, are complicated by the fact that clinicians are still not diagnosing some cases of TRALI, though it’s not clear how many. Dr. Toy says her researchers, through very close monitoring, initially found an incidence of one TRALI case in 3,000 units —a much higher rate than the one-in-5,000-units figure that was accepted before the AABB recommendations were implemented.

A study that Drs. Popovsky and Kopko published seven years ago performed a lookback at recipients of products from a “hot” donor whose plasma had been linked to a death. “We went back and tracked 36 charts traced to her, and we found 15 reactions that were TRALI or TRALI-like. But only two were reported to the transfusion service,” Dr. Popovsky says.

Studies based on random audits have found large numbers of reactions in charts that were not identified as blood reactions or not reported at all. “That suggests that there are more cases of TRALI that are occurring, and we’re simply not finding out about them. So we have a ways to go.”

Moreover, in clinical settings TRALI is often confused with TACO (transfusion-associated circulatory overload). Dr. Toy and others at the University of California at San Francisco, for example, prospectively looked at patients with hypoxemia. “When a lab value was obtained showing low oxygen, they investigated to see how frequently it occurred and was it TRALI or TACO,” Dr. Popovsky says.

“What they showed was the incidence of both TRALI and TACO was high and that often TRALI was misdiagnosed when it was really TACO, or the patient might have both. Because of interest in respiratory problems due to blood, we’re learning about both types of complications.” Until recently, the U.K.’s hemovigilance system didn’t capture TACO, but it has now been added to the transfusion-related events being reported.

“TRALI and TACO are clinically very difficult to differentiate,” Dr. Benjamin points out. “TACO requires some clinical acumen and some laboratory tests can help, but it’s a difficult diagnosis to make.”

However, hemovigilance data in Quebec indicate TACO is a far larger problem than TRALI by an order of magnitude, Dr. AuBuchon notes. “In Quebec, it is the most common cause of transfusion-related death. Based on the data from Quebec—and I don’t know why it would be different from the U.S.— the likelihood of death from TACO is much higher than death from TRALI. So we’re missing it.” However, he adds, TACO is not a problem that blood bankers can address by changing donor screening.

Nevertheless, Dr. Benjamin believes clinician awareness of TRALI has improved significantly over the past six or seven years. “We have seen incremental reports of TRALI almost every year, suggesting that clinicians are getting better at recognizing it.” It used to be you could find TRALI only in the blood bank literature, Dr. Kopko agrees. “Now you see it in pulmonology and anesthesia literature. So I do know they’ve caught on to this and realize it’s a really significant problem.”

Are blood centers doing everything they can to reduce TRALI risk? Dr. Popovsky stresses that the transfusion medicine community did a lot of good in the past 20 years in significantly reducing the risk of infectious diseases through transfusion. “Frankly, I don’t think we get enough credit. But we’ve not done enough to decrease the risk of noninfectious complications. We should not make excuses not to implement the tools we have. Clearly you need enough product on the shelf, but I think there isn’t really a great cause of concern that patients would die for lack of blood.”

A possible future solution may lie in platelet additive solutions, which have been approved in the U.K. and are still pending FDA approval in the United States. “We’re very interested in platelet additive solutions,” Dr. AuBuchon says. “They would allow us to reduce the amount of plasma in each platelet unit to 20 percent or 30 percent and have the remainder of the volume and metabolic substrates for platelets provided by the crystalloid solution.” It’s thought that one reason why red cells are infrequently associated with TRALI is that most of the plasma in the units has been replaced with the red-cell additive solutions, which U.S. blood centers have used for decades, he says.

Dr. Menitove believes platelet additives may help bring TRALI risk down. But, he warns, “There’s still going to be one-third of the suspension volume that contains plasma, and that’s probably at the cusp for not fully mitigating the problems. So I think additive solutions are a good first step; we need to see the next generation of additive solutions that will get us to even less plasma in the components.”

In the U.K., however, a plan to conduct trials of platelet additive solutions has been put on hold. “So far we haven’t felt that additive solution would be a very helpful thing to do,” Dr. Williamson says. In addition, there would be extra cost, “and we can’t fractionate our own plasma, so the plasma that we would take off would just have to be discarded.” However, the National Health Service’s safety committee is going to be considering pathogen inactivation of platelets next year—and that comes with platelet additive solution.

The road ahead could include stricter standards. The AABB and FDA recommendations during the past decade have in many cases been voluntary and up to an individual blood center’s discretion, Dr. Benjamin notes. “I do believe at this point the data on plasma is becoming strong enough that it could be introduced in standards of some sort,” he adds.

TRALI is not going to go away, Dr. Makar emphasizes. “We’re focusing on reducing the risk of TRALI by screening, and making sure that plasma and eventually single-donor platelets are negative for HLA antibodies as far as we can detect, but there are no good screening tests for anti-granulocyte antibodies. They’re comparatively uncommon but they’re out there. So pathologists still need to encourage vigilance on the part of our clinical colleagues as we make an effort to make TRALI less common.”

Still, the past couple of years have been exciting, Dr. Popovsky says. “We have a better understanding of the mechanisms that cause TRALI, and we’re doing something significant about the problem now. We’re seeing the reward for our efforts. All one has to do is look at the recent FDA data and see we did the right thing.”

Related Links

• Reports of TRALI by implicated blood product, FY 2005-FY 2008 (PDF, 1.2 MB)