Transfusion‐related acute lung injury

Abstract

Transfusion-related acute lung injury (TRALI) is best described as a clinical constellation of signs and symptoms including dyspnoea, hypotension and fever with bilateral pulmonary oedema that usually develops within 4 h of a transfusion (33). Patients with TRALI often require respiratory support, but symptoms tend to resolve within 96 h. TRALI is most often associated with transfusion of whole blood, packed red blood cell (pRBCs) and fresh frozen plasma (FFP). There are also reports of TRALI following transfusion of granulocytes (27), cryoprecipitate (35), platelet concentrates (34; 44) and plateletpheresis (12). Infusion of even small volumes of blood can trigger a reaction (4). Estimates of frequency have ranged from 0.014% to 0.02% per unit transfused (33; 1) and from 0.04% to 0.16% per patient transfused (33; 1; 46). Of particular concern is that TRALI has been reported as the second most common cause of fatal transfusion reactions (36). The syndrome is fatal in approximately 5% of cases (47; 33). The first published report of fatal pulmonary oedema related to transfusion was in 1951 (2). Over the subsequent 30 years numerous reports of pulmonary oedema associated with transfusion have appeared in the literature. These cases were assigned various designations including: noncardiogenic pulmonary oedema (45; 6; 9), pulmonary ‘hypersensitivity’ (42; 47) and severe allergic pulmonary oedema (18). The underlying cause of the pulmonary oedema was attributed to incompatibility of an undetermined nature (30), non-HLA leucoagglutinins (42), pulmonary allergic reaction (18; 9; 28) and undefined granulocyte leucoagglutinins (10). The term TRALI was coined by 31). The aetiology of the syndrome was attributed to leucoagglutinating and/or lymphocytotoxic antibodies in the plasma from multiparous donors, directed against the white blood cells of patient recipients. The clinical presentation of TRALI is indistinguishable from adult respiratory distress syndrome (ARDS) (31). Symptoms include dyspnoea, cyanosis, hypotension, fever, chills, cough and production of fluid from the endotracheal tube in intubated patients, along with physical findings of bilateral pulmonary oedema. The symptoms often commence within 1–2 h of transfusion and are usually present by 4–6 h (33). An alternate, less common, presentation consists of mild (22; 45) or no (18) initial symptoms, with the development of full-blown TRALI as late as 2 d following transfusion (22). The severity of symptoms can range from relatively mild to severe, but is primarily related to the degree of hypoxaemia. The syndrome is often associated with significant morbidity. In a series of 36 patients with TRALI (33), all required oxygen support for a mean of 40 h. Mechanical ventilation was required by 72%; TRALI was determined to contribute significantly to mortality in 6%. The development of bilateral pulmonary infiltrates after infusion of blood containing a leucoagglutinin was first described by 4). Pulmonary infiltrates appear at the time of the reaction and resolve within 96 h in about 80% of affected patients (33) (Fig 1). Arterial blood gas values also tend to become altered but return to baseline within this time frame. Infiltrates persist for at least 7 d in the remaining 20%. Persistence of infiltrates has been associated with difficulty weaning patients from mechanical ventilator support (33). The radiographic findings tend to be more remarkable than the physical findings (30; 33). Sequence of chest X-rays taken of a patient with transfusion related acute lung injury (TRALI). Left: normal chest X-ray prior to transfusion; centre: chest X-ray, 2 h after transfusion, showing bilateral pulmonary infiltrates consistent with pulmonary oedema; right: chest X-ray, 48 h after transfusion, showing clearing of pulmonary infiltrates. (Reprinted with permission from 33.) White blood cell antibodies are frequently identified in donor serum as part of a transfusion reaction evaluation following an episode of TRALI. The first documented leucoagglutinin as the cause of a post-transfusion respiratory reaction was reported by 4). A volunteer was administered plasma, from two patients, containing a weak leucoagglutinin; a mild respiratory reaction occurred after both infusions. Whole blood (50 ml) from a patient with hypoplastic anaemia and a strong leucoagglutinin was subsequently given to the same volunteer; a marked respiratory reaction, associated with bilateral pulmonary infiltrates, ensued. In the largest documented series of cases (33), granulocyte antibodies were identified, in the serum of at least one donor, in 89% of the 36 cases. Lymphocyte antibodies were identified in at least one donor in 72%. The blood products implicated were whole blood (21), pRBCs (10) and FFP (five). There are several other small series of cases or case reports in which granulocyte and/or lymphocyte antibodies were identified in at least one donor (13; 45; 42; 6; 10; 5; 31; 15; 22; 21; 1; 14; 17; 23; 34; 44). The exact specificity of the leucocyte antibody involved was determined in only a few cases. These include: donor anti-NA2, recipient NA2 positive (48); donor anti-5b, strongly reactive with recipient granulocytes (26); anti-NB2 in donor, recipient NB2 positive (43); anti-NB1 in donor, incompatible with recipient white blood cells (WBCs) via crossmatch (3) and donor anti-HLA-B35, recipient B35 positive (11). In this last case report an eluate from post-mortem lung tissue was cytotoxic for seven of eight HLA-B35-positive panel lymphocytes and nonreactive for all panel lymphocytes lacking the HLA B35 antigen. Although WBC antibodies are usually found in the donor, TRALI can occasionally occur due to antibodies in the recipient. 47) reported a fatal reaction, associated with the transfusion of pRBCs, in which recipient serum contained lymphocytotoxic antibody specific for HLA-A11 present on the donor's leucocytes. 3) described an episode of TRALI in a patient with granulocyte-agglutinating anti-HLA-A2. The reaction occurred after transfusion of buffy coat poor pRBCs from an HLA-A2-positive donor. Additionally, interdonor TRALI reactions have been reported. 12) reported a case where a plateletpheresis donor possessed lymphocytotoxic antibodies against the HLA-A2 cross-reactive group (CREG) that consists of HLA-A2, A23, A24 and A28. None of these antigens were present in the recipient. However, the A23 and A24 antigens were present in another donor whose RBCs were transfused 24 h prior to the platelets. The patient received three additional HLA-matched platelet transfusions (plasma removed), from the same platelet donor, without a similar reaction. 27) described a near fatal TRALI reaction in a neonate receiving exchange transfusion. Plasma used to reconstitute the RBCs contained granulocyte antibodies. The infant's pulmonary reaction occurred during the subsequent infusion of granulocytes. Granulocyte antibodies were not present in the neonate or mother prior to the exchange transfusion. 44) reported a case of an interdonor TRALI reaction that occurred subsequent to the infusion of a pool of platelet concentrates. The platelet pool was composed of plasma from a single donor, in which platelets from the plasma donor and three other donors were suspended. The plasma donor was strongly positive for anti-HLA-A2 and A28. The recipient patient was negative for the HLA-A2 and A28 antigens. However, the donor of one of the platelet concentrates was positive for HLA-A28. WBCs from this donor were strongly reactive by lymphocytotoxicity and leucoagglutination with serum from the plasma donor. 32) have hypothesized that donor antibodies more commonly cause TRALI than recipient antibodies because the former are able to react with the entire circulating (and marginating) pool of WBCs in the recipient. Antibodies in the recipient have a much smaller pool of donor WBCs in a blood component with which to react. The diagnosis of TRALI is based primarily upon clinical signs and symptoms, not laboratory findings. It is important to determine that the pulmonary oedema is noncardiogenic, because it is treated differently than cardiogenic or volume overload types of pulmonary oedema. The laboratory investigation, although important to confirm the diagnosis of TRALI, is performed at a later date. The presence of noncardiogenic pulmonary oedema following a transfusion should prompt immediate medical treatment of this type of pulmonary oedema with subsequent laboratory confirmation of the presumed TRALI reaction. 32) have suggested a stepwise approach to confirm the diagnosis of TRALI. First, the gender of the donors of all components transfused within the 6 h prior to the reaction should be determined. All female donors are then questioned regarding pregnancy and transfusion history; initially, such donors with four or more pregnancies are tested for HLA and granulocyte antibodies. If negative, donors with one to three pregnancies are then tested. When antibodies are found, a lymphocytotoxicity crossmatch between donor serum and recipient WBCs is performed. If the crossmatch is positive, the diagnosis of TRALI is confirmed. If the crossmatch is negative, TRALI is still presumed in the appropriate clinical setting (see subsequent discussion on pathophysiology regarding another postulated mechanism for TRALI). Pulmonary oedema can be classified into three types (25): cardiogenic, secondary to myocardial or valvular heart disease; overhydration, secondary to excess intake and/or inadequate output of fluid; and noncardiogenic, secondary to increased vascular permeability due to a variety of pathologic, traumatic and infectious causes. TRALI, a form of noncardiogenic pulmonary oedema, is clinically distinguished from other forms of pulmonary oedema based upon normal (9; 10; 31) to decreased (28) pulmonary capillary wedge pressure (PCWP), normal pulmonary artery pressure (10), absence of jugular venous distention (6), absence of murmers or gallops (6), normal cardiac silhouette (10; 48), absence of pulmonary vascular congestion (10) and no evidence of myocardial infarction on EKG (28). Intubated patients who develop TRALI are typically described as producing copious quantities of frothy oedema fluid from the endotracheal tube (9; 31; 16; 17). The protein content of the oedema fluid can be used to help distinguish between types of pulmonary oedema. The protein content of oedema fluid in TRALI is elevated (9; 16). The ratio of protein in oedema fluid to protein in blood is usually >0.7 in noncardiogenic pulmonary oedema and <0.5 in cardiogenic pulmonary oedema (41). The relative increase in protein content in noncardiogenic pulmonary oedema, compared to cardiogenic pulmonary oedema, is thought to be due to increased vascular permeability. In fact, increased vascular permeability was demonstrated in TRALI by 6). Contrast agent was injected into a patient's main pulmonary artery shortly after a TRALI reaction began; the late arterial and venous phases showed extravasation of the contrast agent into alveoli. Corticosteroids, epinephrines and diuretics were traditionally used to treat TRALI. However, 28) described a patient with TRALI who responded poorly to treatment with diuretics. Although the patient had PE, he was not overhydrated. Treatment with diuretics led to decreased PCWP, decreased cardiac output and hypotension. However, subsequently, the patient promptly responded to saline infusion. 16) made a similar observation in a series of eight patients who experienced a total of nine episodes of TRALI after receiving FFP during cardiopulmonary bypass. In the first six patients there was a progressive decrease in PCWP and cardiac output, not improved with an intra-aortic balloon pump or the administration of catecholamines. Three deaths occurred secondary to decreased cardiac output in this group. The last two patients, one of whom had two reactions, were treated with saline infusion to restore adequate left-sided filling pressure and achieve adequate cardiac output. Symptoms resolved rapidly, after saline infusion, in the patient with a single episode of TRALI. The patient who had two episodes responded promptly to treatment with saline infusion the first time. The second episode occurred a few hours later, when another transfusion of FFP was administered. The patient's cardiac output again rapidly improved with saline administration, but he needed ventilatory support for 5 d. 22) and 14) have also published case reports describing the attributes of treatment with respiratory support plus fluid administration versus potentially dangerous treatment with diuretics, antibiotics and invasive procedures. 9) described three cases of TRALI during cardiopulmonary bypass that responded dramatically to infusion of albumin solution. However, treatment with albumin may be difficult to recommend at this time due to a recent meta-analysis that found increased mortality in critically ill patients treated with albumin solution versus crystalloid (Cochrane Injuries Group Albumin Reviewers ( ). Since the pulmonary oedema in TRALI is not related to fluid overload or cardiac dysfunction, but to altered vascular permeability in the lungs, with exudation of fluid and protein into the alveoli, it is logical that maintenance of haemodynamic status is the most beneficial and appropriate therapy. Ventilatory assistance and saline infusion are probably the only therapies that can be recommended, as standard therapy, for the treatment of TRALI. The use of corticosteroids remains controversial, and the use of diuretics may be detrimental. The donor of the implicated blood component is usually multiparous (45; 42; 18; 31; 11). This is primarily related to antibody formation in female donors, due to exposure to paternal leucocyte antigens from the fetus, during pregnancy. 29) demonstrated that leucoagglutinins are present in approximately 18% of parous women. The percentage of women with leucoagglutinins increases with increasing number of pregnancies (29). Payne also demonstrated that 55% of women tested still possessed leucoagglutinins 3 years after initial testing and up to 8 years after the last potential exposure (pregnancy). 7) studied serum from 2313 postpartum women; lymphocytotoxic antibodies were present in 17.2% and granulocyte agglutinating antibodies in 12.6%. For the latter, only two were determined to have an identifiable specificity: one was anti-NA1 and the other anti-NB1. Several approaches to the prevention of TRALI have been offered. 31) and 33) have suggested that blood from implicated donors should only be used as frozen-deglycerolized or washed RBCs. A more recent publication (32) suggested that implicated donors should be told not to donate again. Additionally, it was recommended that blood from unscreened multiparous donors, defined as having three or more pregnancies, should be diverted to recovered plasma and not used as whole blood, FFP, or pheresis platelets. The use of plasma-poor cellular components (platelet concentrates, volume reduced or albumin resuspended platelets and pRBCs stored in protein poor solutions) was neither advocated nor discouraged. This is estimated to affect about 5% of all volunteer blood and component donations. Other authors (34) have suggested that plasma from implicated donors should only be used for fractionation into plasma protein derivatives and all cellular components should be plasma free (i.e. either saline washed or frozen/deglycerolized). Given the rarity of TRALI, a more moderate approach is probably more realistic. Plasma from implicated donors should be diverted for protein fractionation. Transfusion of pRBCs from such donors when preserved in an anticoagulant-preservative solution like AS-2 (e.g. Adsol solution) is probably acceptable due to the small volume of plasma present in this component. The idea of deferring all unscreened multiparous women as blood and pheresis donors is particularly disturbing, considering the ongoing difficulties with blood shortages. Particular caution may need to be taken in cases of patient-directed donation of blood and blood components from relatives, particularly if the donor is the mother of the intended recipient. 5) and 15) reported two cases of TRALI subsequent to donor-specific blood transfusion, in preparation for living-related renal transplantation. In both cases the implicated donor was the patient's mother, whose plasma contained an antibody specific for her child's WBCs. This antibody was presumably acquired during pregnancy. Transfusion in this setting, as well as any transfusion from mother to child, is the perfect setup for TRALI. The mother has already been exposed to the child's WBCs, and has therefore had the opportunity to develop specific antibody to the recipient's paternal WBC antigens, which makes the recipient a prime candidate for TRALI. Theoretically, patient designated donation of blood from a woman to the father of her children could also result in a TRALI reaction, if the donor had formed antibodies to the recipient's leucocytes during pregnancy. Several case reports of TRALI included gross and/or microscopic descriptions of the lungs at autopsy (13; 18; 47; 33; 1; 11; 39). The lungs were typically described as firm, heavy, consolidated and congested (1). Microscopically, the tissue showed diffuse alveolar damage (DAD) with intra-alveolar oedema and haemorrhage, hyaline membrane formation, alveolar cell hypertrophy and scant interstitial inflammation (Fig 2) (39). DAD is a pattern of pulmonary injury that is also seen in ARDS (19). Bronchopneumonia can be superimposed on the DAD if the patient lives for several days after the TRALI reaction (1). the aetiology of TRALI has been attributed primarily to the of usually of donor with WBCs of the patient recipient. When an for WBC antibodies is are identified, in at least one donor, in approximately of cases of TRALI (33). demonstrated that granulocytes to the lungs, when transfused to with granulocyte antibodies. Although this is the of the TRALI it that the presence of leucocyte antibodies can cause pulmonary of and the of antibodies with leucocyte antigens are therefore probably for the respiratory symptoms of TRALI. have TRALI in an lung When the lung is with plasma containing anti-5b, by granulocytes and plasma as a of severe after a of h. The pulmonary oedema is associated with an increase in lung vascular permeability. If granulocytes are or if is not the pulmonary reaction not This of which is not usually thought to occur with and have reported in the of a TRALI reaction. attributed the in case to the presence of a and also a has been reported by 1). Although with subsequent of WBCs in the pulmonary to increased vascular permeability and exudation of fluid and protein into the alveoli, to be the cause of TRALI, two are TRALI reactions not seen when a blood containing leucocyte antibodies is transfused to a patient with the on leucocytes and are leucocyte antibodies not identified in all cases of reported a case of TRALI associated with transfusion of pRBCs from a donor with NB2 antibody into an the implicated donor had and was implicated in a TRALI reaction, even the frequency of NB2 is The authors suggested the patient's her to develop TRALI. 33) first that the of TRALI reactions occur in the setting of This to the that may to TRALI. studied patients with TRALI and patients with or reactions as a group. All patients with TRALI had a These included administration, recent or transfusion. two patients from the group had a a to the aetiology of TRALI. The first consists of a The second is the infusion of in stored blood This is primarily based upon the observation that stored cellular blood at the time of a that cell This is not present in fresh cellular blood components or blood These demonstrated that there was significantly more present in post-transfusion from the patients who had TRALI reactions compared to and and post-transfusion from the patients with or transfusion reactions (39). This group has also a mechanism of TRALI in a lung The were with to lungs were subsequently with either fresh plasma, plasma from stored RBCs on the of or the of from plasma or injury was not in lungs with fresh plasma or plasma from it was in lungs with plasma from pRBCs, from plasma or This as the main cause of TRALI reactions to into a that about TRALI and has The consistent presence of leucocyte antibodies is in TRALI is to be The presence of in post-transfusion TRALI may well be not not significantly in fresh plasma or FFP, the that TRALI may occur secondary to infusion of FFP 16; 33). If the required two a significant and transfusion of blood products at or near TRALI should be an common in and In fact, it probably on a in most medical An or may the of TRALI. The first is a in the It is the that most TRALI reactions in either an unit or an The second is the transfusion of a blood leucocyte antibody directed against the patient's WBCs, or (3) leucocyte antibody and This antibodies are not found in all cases of TRALI. It also TRALI has occurred secondary to the administration of FFP, although are not present in It may also leucocyte antibodies in a donor blood component not cause TRALI at the frequency A may need to be for a reaction to If antibody is not the additional presence of may be needed for a reaction to it also all components transfused at or near to significantly ill patients, not cause TRALI. There may not be present to cause a reaction in the absence of donor leucocyte antibody to a patient's WBCs. TRALI is a clinical constellation of signs and symptoms, related to noncardiogenic pulmonary oedema, that subsequent to transfusion. Antibodies to usually of donor are identified in the of presence is Treatment should be laboratory which typically at a later date. Treatment should be when TRALI is first and consists of fluid support to blood pressure and cardiac output plus ventilatory have no in the treatment of use may be detrimental.