© 2007 American Thoracic Society doi: 10.1164/rccm.200708-1169ED
TRALIA New Case Definition, a New Epidemic?University of Washington and Harborview Medical Center, Seattle, Washington
University of Toronto and St. Michael's Hospital, Toronto, Ontario, Canada Transfusion-related acute lung injury (TRALI) represents acute lung injury (ALI) after the transfusion of one or more plasma-containing blood products and is now the leading cause of transfusion-related morbidity and mortality (1). The incidence of TRALI is frequently reported as 1 occurrence for every 5,000 blood component transfusions (2, 3); however, this is thought to be a vast underestimate of the true incidence, resulting from lack of recognition or underreporting. Two hypotheses have emerged to explain the mechanisms responsible for TRALI. One involves an interaction between passively transferred donor human leukocyte antigen (HLA) antibodies and recipient HLA antigens (3, 4). The cognate antibody–antigen interaction leads to leukocyte activation and subsequent lung injury. Alternatively, a "two hit" hypothesis has been proposed in which biologically active substances, such as lysophosphatidylcholine species of bioactive lipids, accumulate in the red cells as the blood is stored and are transfused passively (5). These inflammatory lipids can subsequently induce neutrophil priming capable of producing lung injury in the setting of systemic inflammation from another insult, such as trauma or sepsis (6, 7). Despite its frequency and morbidity, we have an incomplete understanding of the epidemiology and basic mechanisms of TRALI. Like other critical illness syndromes, this is due, in part, to unreliable and inconsistent definitions. Most TRALI studies exclude cases in which another ALI risk factor is present. While this is a justifiable decision, such a restrictive definition will prevent any evaluation of the two-hit hypothesis because patients primed by systemic inflammation would be excluded. To address the lack of a common TRALI definition, the National Heart, Lung, and Blood Institute's Working Group on TRALI published a consensus definition in 2005 (8). According to this report, TRALI is now recognized as new-onset ALI occurring within 6 hours of transfusion of a plasma-containing blood product. ALI is largely defined according to the 1994 ALI/ARDS North American–European consensus conference definition (9). Notably, unlike the prior definition, the new consensus definition allows for a possible diagnosis of TRALI if other classic ALI risks are present yet the clinical setting makes TRALI likely. In this issue of the Journal (pp. 886–891), Gajic and coworkers (10) evaluated the incidence of TRALI among a cohort of transfused medical intensive care unit (ICU) patients using the 2005 TRALI definition to better understand the implications of this case definition. By using electronic surveillance methods, coupled with a clinical expert review panel, they capture both suspected TRALI cases (with no identifiable preexisting ALI risk factor) and possible TRALI cases (with an additional ALI clinical risk factor present). The authors report 74 new TRALI cases arising from 901 observed patients and 6,558 associated blood products (inclusive of packed red blood cells, platelets, and fresh frozen plasma) transfused. This incidence is about 50 times higher than previous estimates. However, it is important to place this finding in context. This study was conducted in a single-center medical ICU with a generous proportion of patients having gastrointestinal (GI) and/or liver disease. Of their TRALI cases, 62% had another ALI risk factor, thus highlighting the close interaction between transfusion, classic ALI risk factors, and the subsequent development of ALI. This finding supports more contemporary epidemiologic studies indicating that transfusion of even a single unit of blood to critically ill patients with other preexisting ALI risks is independently associated with a greater risk of subsequent ALI and death (11, 12). Although the TRALI definition allows the investigators to explore the two-hit hypothesis, teasing out the independent effect of blood products from the indications for blood products, which may be multiple, in patients at risk for ALI remains a challenge. The authors concede that potential misclassification of cases (i.e., ALI brought on by the traditional ALI risk factor and not transfusion) as well as some degree of residual bias, including indication bias, may be present. This study used a nested case-control design well suited to evaluating the interaction between clinical and biologic factors in patients as well as the donor blood products. The authors perform a multivariable logistic regression analysis showing that both specific clinical donor characteristics (donor gender and parity) and blood product characteristics (HLA class II antibodies and bioactive lipids) have independent effects on the development of ALI. This supports the notion that soluble bioactive agents in the plasma compartment, comprising what is collectively referred to as the "storage lesion," are associated with the development of TRALI. Interestingly, the age of blood did not differ between cases and control subjects. Current data support the idea that soluble bioactive agents, including inflammatory lipids, accumulate in the plasma compartment as blood products are stored (5, 13, 14). Observational data have linked the storage time of red cell units with multiple organ failure, death, and other adverse clinical outcomes (15). By analyzing the storage time as the median duration of blood storage time, the authors may have diluted the effect of one older blood unit on TRALI development among the several transfused. Furthermore, the phenomenon of transfusion-related immunomodulation (TRIM) might predispose to infection, creating an associative link between transfusion and ALI (through sepsis) (16). In a randomized controlled trial of trauma patients, we found no advantage to leukoreduction in preventing infection (17); however, other biologically active modifiers in stored blood could be responsible for TRIM. Gajic and colleagues' study does not provide us with information to temporally link transfusions with the development of sepsis, but sepsis was the most common risk factor associated with ALI and might confound the relationship between transfusion and lung injury. Finally, it is important to point out that around 50% of the associated units in TRALI cases and over 25% of the associated units in control subjects did not have analyses for the stated potentially bioactive factors, which could introduce bias into these biomarker estimates. Gajic and colleagues have provided us with data that draw attention to the frequency with which TRALI occurs in the medical ICU. If the true incidence of this syndrome is even half of that estimated by this study, the associated morbidity and mortality occurring in the more than 4 million estimated recipients of blood products each year is enormous. Future studies should focus on other at-risk cohorts, including those with severe trauma, sepsis, and GI-related hemorrhage. These analyses will serve to further confirm and define the epidemiology of this syndrome and to establish the etiologic role of storage age, HLA antibodies, bioactive lipids, and other potential components of the storage lesion. In addition, the effect of alternative blood processing strategies, including red blood cell washing and targeted donation based on gender and parity, need to be investigated. Whether TRALI carries unique mechanisms of lung injury, setting it apart from those pathways responsible for ALI attributable to other risks, is unknown. This new case definition and the resultant higher disease incidence will provide the impetus to more thoroughly evaluate the relationship between transfusion and lung injury. FOOTNOTES Conflict of Interest Statement: Neither author has a financial relationship with a commercial entity that has an interest in the subject of this manuscript. REFERENCES
Related articles in AJRCCM:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
