On the trail of TRALI, the latest Tx villain
William Check, PhD
It would not be surprising if physicians and administrators
charged with protecting the nation’s blood supply felt like they were facing the typical endless roster of villains in a video game: As soon as they eliminate one threat to transfusion safety, another pops up. With blood screening tests now in place to block transmission of HIV and HCV, the danger from those viruses has declined to a virtually undetectable level. "Transmission of HIV and HCV through blood is so rare now that we can’t measure it," says Patricia Kopko, MD, associate medical director of BloodSource (formerly Sacramento Blood Centers) and assistant clinical professor of medical pathology at the University of California at Davis. "It can only be calculated from modeling."
Has vanquishing two killer viruses earned blood bankers a rest? Fat chance. As if to prove that eternal vigilance is the price of safety, removing infections has unmasked yet another villain: transfusion-related acute lung injury, or TRALI. During the 1990s, TRALI was the third-leading cause of transfusion-related fatalities, behind ABO incompatibility reactions and bacterial contamination. However, it is now the No. 1 killer among blood products, according to data from the Food and Drug Administration from the past three years. Specialists in this field attribute TRALI’s rise to the top of the most-wanted list not to increased incidence, but to increased awareness and recognition. "TRALI has been a hot topic for the last couple of years," Dr. Kopko says. "We used to get maybe two cases per year. Already this year we have had five probable cases reported." Current estimates are that plasma-rich products cause perhaps 1,000 severe TRALI reactions per year, including 100 to 200 deaths.
Increased appreciation of TRALI’s importance led Morris Blajchman, MD, professor of molecular medicine and pathology at McMaster University and head of transfusion medicine at McMaster Hospitals, to organize a consensus conference this past April on the condition. "My biggest motivation was the growing realization that this is a very serious problem associated with blood transfusion," Dr. Blajchman says, "and that we really understand very little about it."
Dr. Blajchman’s profession of ignorance is not false modesty. Consider the
literature review that he and transfusion medicine fellow Kathryn Webert, MD,
published just prior to the conference (Webert KE, Blajchman MA. Transfusion
Medicine Reviews. 2003; 17: 252-262). In this review is a table of 10 "Unanswered
Questions and Controversial Issues" regarding TRALI. They include: definition,
diagnostic criteria, prevalence, risk factors, pathophysiological mechanisms,
workup of suspected cases, treatment, management of donors implicated in TRALI
episodes, and prevention.
In fact, considerable data exist on these topics, but for the most part they
are incomplete and inconclusive. So what was there to say at the conference?
While it didn’t provide breakthroughs or solve outstanding problems, there was
agreement on certain issues and on a research agenda. "I think it will serve
as a starting point as to how we can become serious about recognizing, diagnosing,
and treating TRALI," Dr. Blajchman says. (According to Dr. Blajchman, the consensus
panel statement will appear in Transfusion and the proceedings in Transfusion
Medicine Reviews in early 2005. An abbreviated statement will appear on
transfusion-related Web sites in July.)
"What wasn’t new [at the conference]—I think that’s more the issue,"
says James AuBuchon, MD, E. Elizabeth French professor and chair of pathology,
Dartmouth Hitchcock Medical Center, and chair of the CAP Transfusion Medicine
Resource Committee. "It’s not new that we don’t have a foolproof or even better
means of preventing TRALI. We are still stuck having to offer as our primary
approach early detection and supportive therapy rather than true prevention."
One helpful outcome will be a
new definition. "At least people will be gathering cases in the same way," Dr. AuBuchon says. While the definition is not finalized, it will most likely emphasize symptoms: dyspnea and hypoxemia with fever and possibly hypotension, with bilateral pulmonary edema in severe cases, occurring from the start of transfusion to four to six hours later. To require demonstration of antileukocyte antibodies may be premature. "There remains viable more than one theory as to the cause of TRALI," Dr. AuBuchon says. "Some cases are caused by antileukocyte antibodies, while others have no apparent cause."
Dr. Blajchman puts a more positive spin on this observation. "One of the most important ideas that came out at this meeting," he says, "is that TRALI is not a single entity. It is likely that there is more than one pathophysiologic mechanism. An important one is antibodies to leukocytes. But we have certainly seen classic TRALI episodes in recipients of blood that came from male donors who don’t have antibodies and have never been transfused." Moreover, both donor and recipient factors seem to be important. "Not everybody who receives a unit containing antileukocyte antibodies is at risk," Dr. Blajchman says. "Recipient factors are something we know nothing about."
Recipient factors figure in the two-event hypothesis, which postulates a vulnerable
donor exposed to plasma containing a provocative agent. It was first proposed
in 1990 by researchers at the University of Minnesota Medical School (Van Buren
NL, et al. Transfusion. 1990; 30: 42-45) and later demonstrated to
be correct by Christopher Silliman, MD, PhD, associate professor of pediatrics
and surgery at the University of Colorado School of Medicine and associate medical
director of Bonfils Blood Center, Denver, and his colleagues (Silliman CC, et
al. Transfusion. 1997;37:719-726). Dr. Silliman also proposed at that
time that biologically active lipids could act as TRALI triggers. He explains
that he derived both ideas by analogy to adult respiratory distress syndrome,
or ARDS, from which TRALI is clinically indistinguishable. "The two-event pathogenesis
of ARDS was shown in animal models to result from two lipids," he says, "endotoxin
and platelet activating factor."
"I think that the two-hit explanation is probably right," says Carol Marshall, MD, associate clinical professor of pathology and associate medical director of the blood bank at the University of California at Davis Medical Center. Such patient factors as sepsis, recent surgery, or infection could predispose to TRALI. "I wonder if there is something yet to be found, some particular cytokine, that is triggering this reaction," Dr. Marshall muses.
Antileukocyte antibodies were fingered as a cause of TRALI in two early papers
in which the syndrome was first described and named (Popovsky MA, et al. Am
Rev Respir Dis. 1983;128:185-189; Popovsky MA, Moore SB. Transfusion.
1985; 25: 573-577). Addressing the relative roles of antibodies and lipids in
TRALI, Mark Popovsky, MD, first author on those papers, says, "Most people now
think that both forms of pathogenesis exist. I think most people believe the
antibody-mediated route is more prevalent. But the court is still out. We need
more data." Dr. Popovsky is senior vice president and corporate medical director
at Haemonetics Corp., Braintree, Mass., and associate professor of pathology
at Harvard Medical School.
As for prevention, Dr. AuBuchon says, "The conference did not come out with a set of clear directions that we should take nor a mandate to implement a particular action that is very different from what we have been doing in the past." Many blood banks now defer donors whose plasma has been associated with a case of TRALI. Dr. AuBuchon calls this approach "clearly necessary but ineffective. It’s closing the barn door after the horse has escaped," he says. "It does not materially improve transfusion safety for recipients."
In the United Kingdom, the National Blood Service has mandated that fresh frozen plasma from male donors be used for transfusion while FFP from female donors is diverted for fractionation. Blood donations from women are not excluded from platelets or red blood cells. "They decided they had enough information correlating plasma from multiparous donors and TRALI cases, so they took action," says Dr. Popovsky. (Antileukocyte antibodies are present at an elevated rate in multiparous women, about 20 percent.) He is paying close attention to the British experiment. "If they are right," he says, "there should be a dramatic decrease in the number of TRALI cases reported in the next couple of years. I would wager that you will see such an impact."
More drastic steps have been proposed, such as deferring all multiparous females. Dr. Kopko’s response to this idea is blunt: "Do you want to have any blood?" With this approach, about one-third of donors would be lost. "We would have a blood shortage in this country the like of which we have never seen," Dr. Kopko says. "An extreme deferral policy is not justified at this time, especially given the uncertainties about TRALI."
When Dr. Popovsky, then at the Mayo Clinic,
and his colleague Breanndan Moore, MD, reported the first series of TRALI cases, it was made possible by Mayo having a specific transfusion team. In their 1983 article they described a cluster of five patients who had the same symptom complex over a period of several months. In that initial series antileukocyte antibodies were found in all five donors, and corresponding HLA antigens were found in three recipients. The series was expanded to 36 patients in the 1985 article, in which they derived the widely accepted estimate that TRALI occurs in one in 5,000 transfusions.
"We believed that this was a form of lung injury that was precipitated by transfusion but that was in the family of lung injury that ultimately led to ARDS," Dr. Popovsky says. The pathology of TRALI is similar to that of ARDS. But ARDS due to sepsis, for instance, has a much higher fatality rate, 50 percent in the 1980s and even now about 30 percent, whereas the mortality rate for TRALI is much lower, five percent to 10 percent. Of the 36 cases in the 1985 article, 80 percent resolved within 96 hours. This prognostic difference makes it important to distinguish TRALI from lung injury due to other causes.
It turned out that TRALI was not new. "When Dr. Moore and I did a search of the literature," Dr. Popovsky says, "we were convinced this was not a new entity and that it had been observed since 1951, but always under different names and always just a few cases. There had been no common linkage or descriptor that would provide a unifying concept." Drs. Popovsky and Moore characterized the reaction and provided a descriptive name.
So why did it take another 15 years for TRALI to become widely recognized? "When Popovsky and Moore described this reaction, it was the height of the AIDS crisis," Dr. Kopko points out. "There were much bigger problems in transfusion medicine at that time."
She adds, "Solid-phase technology to adequately identify antibodies involved in TRALI has only become available in the last five years or so." Her laboratory uses a flow cytometric technique. ELISA-based techniques are also available. "Both are less prone to ambiguity of interpretation than cellular assays," Dr. Kopko says.
On the positive side, Dr. Popovsky says, "Publications on TRALI began to appear from all over the world in the early 1990s. We saw that it did not occur at just one medical center. It was seen in all developed and even developing countries. That gave credence that this was a real problem."
A further wakeup call came from the FDA. "FDA was concerned about the frequency of reports [of TRALI deaths]," Dr. Popovsky says. "They organized a workshop at a blood products advisory committee meeting." The minutes of the workshop went onto the FDA’s Web site, and in 2001 the agency issued an advisory on TRALI that went to every physician in the U.S., greatly increasing the visibility of this condition.
The importance of recognizing TRALI was underlined by a 2002 lookback investigation
among 54 people who had received blood products from a frequent plasma donor
whose blood product, containing granulocyte 5b antibody, was implicated in a
fatal case of TRALI (Kopko PM, et al. JAMA. 2002; 287: 1968-1971).
Among 36 recipients with complete charts, 13 (36 percent) had had a total of
15 TRALI reactions, showing that donors whose plasma is implicated in one episode
and that contains antileukocyte antibodies may pose "a future transfusion hazard."
Half of the reactions were classified as severe, meaning the patient had evidence
of acute pulmonary edema or needed mechanical ventilation, or both. Only seven
of the 15 reactions had been reported, leading the authors to conclude that
TRALI is "underdiagnosed and underreported."
(Paradoxically, five of the seven mild/ moderate reactions had been reported, but only two of the eight severe reactions. Dr. Kopko suggests that mild reactions were more likely to be reported because they can look like a mild transfusion reaction called febrile nonhemolytic transfusion reaction, or FNHTR. "They teach you about FNHTR in med school," she says. "Anytime you see a one to three degree C temperature rise you are supposed to stop the transfusion and report it." More severe cases of TRALI, on the other hand, are easily confused with nontransfusion causes of respiratory distress.)
Dr. AuBuchon concurs that TRALI is underreported. "I have been here for 14 years," he says. "In that time we would have transfused about 150,000 units of red cells plus 15,000 apheresis platelet units. But we have only had eight cases that we could call TRALI." Statistics suggest that three to four times that number would be expected.
Further evidence supports the idea that one donor’s plasma can cause multiple
cases of TRALI. Says Dr. Kopko, "Following our paper, there was a letter to
JAMA reporting a lookback investigation in which 15 percent of recipients
got TRALI from the same donor." In that case the donor had multi-specificity
antibody, which reacted with basically every cell on the panel.
At least one lookback study found no further cases from implicated plasma (Win
N, et al. Transfusion Medicine. 2002;12:387-389). Asked about the difference
between this lookback and the one she participated in, Dr. Kopko says, "I think
all antibodies are not created equal. I suspect that the one we did the lookback
on, granulocyte 5b, is a really bad actor." She notes two additional fatalities
from this class of antibody (Davoren A, et al. Transfusion. 2003;43:641-645).
Possibly the propensity of an antibody to cause TRALI depends on the prevalence
of its corresponding antigen in the population: Granulocyte 5b antigen is present
in more than 90 percent of Caucasians.
Dr. Popovsky says that unpublished data from the original Mayo study support the existence of "hot" donors. He reviewed charts from one woman who had had 11 pregnancies and who had been a donor for many years. "Nursing or physician notes described respiratory distress in many recipients," he says. "So this donor was causing TRALI, but people didn’t make the connection at the time."
Getting clinicians to "make the connection"
between acute respiratory distress and transfusion has been a challenge. One reason is that, among the pantheon of foes, TRALI is more like wily Odysseus than slashing Achilles. The differential diagnosis of acute dyspnea includes pneumonia, sepsis, ARDS, and heart failure, in addition to TRALI. "In an ill patient especially," Dr. Marshall says, "it can be hard to tell when you are dealing with TRALI." She says it is important to think about TRALI even if you only see a febrile or mild reaction if it is associated with hypoxia or dyspnea and occurs within a few hours after starting a transfusion. "Patients can become volume overloaded by transfusion," Dr. Marshall says. "I find fever to be helpful in the diagnosis. Fever is not usually present in patients with cardiac problems."
Dr. Kopko says the case that provoked their interest in TRALI resulted from misdiagnosis. Forty-five minutes into a plasma transfusion, the patient was found in respiratory distress on the verge of complete apnea. She was intubated and put into the ICU. "Only 200 mL of plasma had been given," Dr. Kopko says, "yet the resident treated her for fluid overload. This is not the best treatment for TRALI." The next morning the pulmonary ICU attending came in and said to the resident, "You do realize that this is a case of TRALI?" The resident had not heard of it.
Even now, Dr. Kopko says, "There is not enough or adequate clinical recognition of the disorder. Unless you are a blood banker, you probably don’t know very much about TRALI. Surgeons and anesthesiologists are there where it occurs; they see patients with acute lung injury. TRALI might not be the last thing they think of, but often it isn’t the first."
"We haven’t adequately educated our physician colleagues," Dr. Blajchman concurs. "Often people outside transfusion medicine don’t know about it." He emphasizes that TRALI can look like heart failure on x-ray and pulmonary findings, but needs to be treated differently.
"We have been out educating whoever will let us come and stand on our soapbox,"
Dr. Kopko says. Publication of the JAMA paper gave her group increased
access to practice meetings, grand rounds, and medical staff meetings. "When
we started," she says, "we were doing a couple of sessions each month." Her
preferred target groups are anesthesiologists and critical care and hematology/oncology
Dr. Marshall, a coauthor on the JAMA article, has also conducted many
educational programs. She targeted internal medicine, hem/onc, and anesthesia.
"I’ve also talked more informally with critical care people," she says. In addition,
her group did a video for nursing inservice that was provided to a variety of
hospital committees on which nurses sit. "I understand the video is shown to
all new nurses," she says.
Dr. AuBuchon has taken similar steps, focusing on nursing staff and anesthesiologists. "Those are the people who are likely to be standing next to the patient during a transfusion," he says. "We want them to recognize symptoms that we would like to hear about in the transfusion service laboratory, so we can direct patient care and workup." On the other hand, he adds, "I am not sure that trying to raise the level of consciousness about TRALI among surgeons and hematologists will be effective because they have so much else to pay attention to."
Dr. AuBuchon zeroes in on nurses, to the extent of aiming to remove physicians from the reporting circuit. "A physician cannot order that a transfusion reaction not be reported," he says. "It doesn’t matter that there might be some other presumed reason for the symptoms. It doesn’t matter that the physician feels the reaction is insignificant and unrelated to the transfusion. The nurse must report the reaction." Transfusion nurses who see respiratory distress are not expected to diagnose TRALI. "But we do depend on them to recognize that symptoms are occurring, stop the transfusion, and report to us promptly."
Dr. Blajchman encountered an incident that encourages him about the value of education. "In our jurisdiction a patient had a severe event following a transfusion," Dr. Blajchman relates. Dr. Blajchman was called, and he diagnosed TRALI. Six months later the same physician who missed the diagnosis in the first case readily diagnosed a subsequent case. "Having the experience made him familiar with TRALI, and he came up with the correct diagnosis on his own the next time," Dr. Blajchman says.
"Awareness of TRALI has definitely increased," Dr. Popovsky says. "Ten years ago when I spoke in hospitals, doctors were often hearing about it for the first time. Now many clinicians have at least an awareness of the problem."
Broader evidence for increasing awareness of TRALI comes from FDA data presented at the Toronto conference. Until 2000, the three leading causes of transfusion-related fatalities in the U.S. were ABO incompatibility, bacterial contamination and sepsis, and TRALI, in that order. In 2001 TRALI became the leading cause of transfusion-related mortality. For the three-year period from 2001 to 2003, TRALI was responsible for 16.3 percent of reported transfusion-related deaths, ABO hemolytic reactions for 14.3 percent, and bacteria for 14.1 percent. "So in the last three years, TRALI has become the leading cause of transfusion-related fatalities based on reporting to FDA," Dr. Blajchman concludes. "And it’s still underreported."
When TRALI is reported, a workup of likely
donors is indicated. Such workups can be expensive. "I have heard people say they don’t want to report it because the blood center is going to charge them so much money to work it up," Dr. Kopko says. In the lookback study, workups were paid for with grant money from the National Blood Foundation. In daily life, Dr. Kopko says, "We have come to an understanding with our hospitals—we pay for donor workup, they pay for patient workup." In the interest of selective testing, only the unit(s) transfused right before the reaction are worked up. "We have found that for the most part the implicated unit is the unit hanging at the time of the reaction or its immediate predecessor," Dr. Kopko says. Two other ways of prioritizing are to test female donors before males and to test high-plasma volume units first (for example, FFP before RBCs).
When a donor has been identified whose plasma contains antibody corresponding to an antigen present in the recipient, deferring that donor is justified, Dr. Kopko says. "They don’t like it," she says. "It makes them very unhappy. But given the couple of lookbacks that have been published that show it can happen again, we feel an ethical obligation to defer those people."
(Even the lookback itself has to be handled "delicately," Dr. Aubuchon notes. "Donors do not like feeling that they have harmed recipients. After all, they donated to help people. Sometimes blood bankers have to be psychiatrists also.")
Dr. Popovsky has raised the possibility of excluding plasma from female donors with HLA or leukocyte antibodies. "I have put that into articles to be controversial," he says, "to make people think." He does not advocate such a measure. "You can have the purest blood in the world, but if there are only a few units it won’t help," he says. "We need to balance blood availability with safety." He thinks a better strategy would be one like the U.K.’s, in which plasma from multiparous donors isn’t transfused but goes into derivatives. In practice, this would need to be done with all female donors. "The problem is reliability of histories," Dr. Popovsky says. You end up asking women if they have ever had an abortion or miscarriage. "You don’t want to offend donors," he says.
But American blood bankers are not ready to adopt the U.K policy. At the consensus conference, Dr. AuBuchon says, "The panel felt there was insufficient evidence to recommend that course of action."
There has been discussion of screening donations from women (or all donations) for the presence of antibodies. "Right now there doesn’t appear to be a simple way that this can be accomplished," Dr. AuBuchon says. "There is not firm agreement on which antigen specificity should be sought." Dr. Kopko notes that there is not even a commercial test for the strongly implicated granulocyte 5b antibody. "We send samples to the American Red Cross Neutrophil Serology Laboratory in St. Paul," she says. "It takes two to three weeks to get data back." Still, she calls screening "worth examining. It might be possible in the next couple of years."
Dr. Popovsky raises the possibility of washing RBCs and platelets to reduce plasma volume. "The technology is available," he says. As for expense, he says, "It would add cost, but if you can prevent an ICU visit or a patient being on a ventilator or even death, I suspect that would be a good return on investment." He says washing would add "dollars per intervention, not thousands of dollars." In Dr. AuBuchon’s view, however, "Washing RBCs would probably work, but it is just not feasible."
Based on his two-event hypothesis, Dr. Silliman proposes washing for selected recipients. "If someone is sick and has a scheduled transfusion," he proposes, "wash it on a machine for $30." Dr. Silliman also recommends using younger products, under 10 days for RBCs and under two days for apheresis platelets, based on his data suggesting that bioactive lipids, which accumulate with age, are responsible for much TRALI.
Many confusing aspects of TRALI—definition, diagnosis,
management of donors, prevention—are confounded by the unsatisfactory
understanding of pathogenesis. A major controversy concerns antibodies versus
lipids as causative agents. In one study, Dr. Kopko and colleagues found antigen-antibody
correlation in 15 (87.5 percent) of 16 cases (Kopko PM, et al. Transfusion.
2003;43:177- 184). At the other extreme, Dr. Silliman and colleagues found antileukocyte
antibodies in only 3.6 percent of 51 cases (Silliman CC, et al. Blood.
2003;101: 454-462). "I think it would be premature to say you always have to
have antibody," Dr. Kopko says. She calls Dr. Silliman’s group’s results "bothersome."
"A lot of people look at our work in Sacramento and Silliman’s group’s work and say it can’t be both," she continues. "I think it can be both." At the Toronto conference Dr. Kopko pointed out differences in transfusion practices between the study populations. Dr. Silliman’s study associating TRALI mainly with platelet concentrate transfusion was done in Canada. "We haven’t used platelet concentrates in our service area in over a decade," Dr. Kopko says. "And we leukoreduce everything. So just because I am not finding what they found does not in any way invalidate their results."
"We have worked up several cases where we can’t find antibodies and are sure the diagnosis is TRALI," Dr. Marshall says. "Maybe Dr. Silliman is right. Maybe some cases of TRALI are related to bioactive lipids." She adds, "Lipids can only be part of the story, though. If it were just lipids, male donors would be as likely to cause TRALI as female donors."
"I think both antibodies and lipids can do it," Dr. Silliman agrees. One of his criteria for the ability to provoke TRALI is the ability to prime neutrophils in vitro. Dr. Silliman, working with Dr. Kopko, showed that 5b antibodies work the same way that lipids do to prime neutrophils. (These results have been submitted to the upcoming American Association of Blood Banks meeting.)
Dr. Silliman’s second innovation, emphasis on recipient factors, has more widespread acceptance. "Surgery, infection, or massive transfusion represents the first event," he says, "which causes neutrophil priming to release chemokines." In his 1997 study all patients with TRALI had either acute active infection or recent surgery within 24 hours, or were massively transfused. Only two patients in the control group had one of these factors. He asserts that all patients in Dr. Popovsky’s original study were within 24 hours’ postop.
To show that nothing about TRALI is simple, Dr. AuBuchon describes a recent
case in which a young man with aplastic anemia who was otherwise healthy received
an RBC transfusion. He began feeling ill five minutes into the transfusion and
began coughing at 10 minutes. At this point the nurse stopped the transfusion.
"The patient was known to be HLA A3 positive," Dr. AuBuchon says, "and the unit
was found to have anti-HLA A3. Would this have turned into a clinically significant
case of TRALI?" he asks. "I don’t know, but I am glad the transfusion nurse
picked up the symptoms." Here is a patient who would not be thought to have
any significant risk factors—no recent surgery, no infection, and otherwise
healthy. He had only mild symptoms, but only a very small portion of the unit
"We need to be able to identify patients at risk and perhaps handle them differently," Dr. Popovsky says. "I think that will be one recommendation that comes out of the conference."
Addressing our ignorance about patient risk factors for TRALI, Dr. Kopko says, "How many Tylenol does it take to kill your liver? It depends on how much whiskey you have been drinking. With TRALI, we don’t even know what the whiskey is.
"We still don’t understand one of the biggest things," she says. "Why is it that when a donor has an antibody and the recipient has the corresponding antigen, not every recipient gets TRALI? We clearly showed in our lookback study that only 36 percent of people who could have gotten TRALI had a reaction. Yet 5b antigen is present in greater than 90 percent of people. So there is a piece missing."
Indeed, there are many pieces missing from the TRALI profile. Now that it has come out of the shadows, perhaps that profile can finally be filled in.
William Check is a medical writer in Wilmette, Ill.