Acute lung injury after transfusion - Complications of Transfusions

Practical Transfusion Medicine 4th Ed.

9. Acute lung injury after transfusion

Steven H. Kleinman¹ & Ram Kakaiya²

¹University of British Columbia, Victoria, British Columbia, Canada

²Life Source Blood Services, Rosemont Illinois, USA

Definition

The clinical syndrome of transfusion-related acute lung injury (TRALI) is characterised by acute onset of respiratory distress during or within 6 hours of transfusion, associated with oxygen desaturation (hypoxemia) and bilateral lung infiltrates, without evidence for left atrial hypertension or circulatory overload. However, differentiation of TRALI from circulatory overload can be difficult, if not impossible. No specific treatment for TRALI exists; in 90% of the cases, the patient recovers completely within 96 hours but the remaining 10% of cases are fatal.

Incidence

Haemovigilance data establish that TRALI is the number one cause of acute mortality from transfusion. In the USA, reports of transfusion-related fatalities to the Food and Drug Administration (FDA) indicate that TRALI has been the number one cause of fatalities from 2005 to 2010 [1]. In the UK, the Serious Hazards of Transfusion (SHOT) reported 36 cases of TRALI in 2003. This annual number decreased in subsequent years, coincident with the introduction of risk mitigation strategies; the 2010 annual SHOT report indicates only 15 reported cases [2].

There is a consensus that TRALI is both underrecognized and underreported; thus, the precise incidence of fatal plus nonfatal TRALI is unknown. An overall risk of approximately 1:5000 transfused units in the general hospital population was reported in 1985 and this incidence number is frequently cited in the literature [3]. In research settings, computer-generated automatic alerts for respiratory distress after transfusion may improve detection of TRALI cases [4]. In a case control study of 89 TRALI cases at UCSF and Mayo Clinic, the investigators identified the following independent patient risk factors for TRALI: chronic alcohol abuse, current cigarette use, pre-existing shock, positive fluid balance, peak airway pressure greater than 30 cm H₂O if mechanically ventilated before transfusion, liver surgery (mainly transplantation) and elevated pretransfusion levels of interleukim-8 [5]. Acute lung injury (ALI) due to other causes is common in critically ill medical patients whether or not they are transfused; however, its frequency is increased in transfused patients (up to 40–45% in some studies) and shows a dose–response relationship [6,7]. This points out the difficulty in determining whether ALI in patients with predisposing ALI risk factors is indeed TRALI or is instead due to other etiologies.

Pulmonary dysfunction that is not severe enough to meet the definition of TRALI may also occur following transfusion. This is consistent with a threshold model of TRALI in which the strength of the mediators (antibodies, activated lipids) may determine the severity of symptoms [8]. Some of these cases are classified as transfusion-associated dyspnea (TAD).

Clinical manifestations

Males and females are equally affected. Most cases occur in adults. Previous transfusion history is unremarkable and recurrent TRALI is extremely rare. At least one case of TRALI from an autologous transfusion has been recorded. A few cases have occurred in children. Directed donations from mother to child can cause TRALI due to maternal leucocyte antibodies against the child's leucocyte antigens.

The onset of TRALI is often quite dramatic, with symptoms occurring either during the transfusion or usually within 2 hours (but this can be up to 6 hours) of its completion. Some investigators have also reported that TRALI can rarely have a delayed onset (>6 hours after transfusion) [9]. The syndrome manifests as acute respiratory distress syndrome (ARDS) or as noncardiogenic pulmonary oedema, and is characterized by acute onset of respiratory distress with dyspnoea, tachypnea and oxygen desaturation. The patient may appear cyanotic and may develop hypotension or hypertension. Oxygen desaturation is often severe, requiring mechanical ventilation in 70% of cases. Mild cases of respiratory distress that are unaccompanied by hypoxia, do not require any oxygen administration or resolve quickly do not fit the diagnostic criteria for TRALI. Some patients with TRALI experience a low grade fever for several hours. Symptoms and signs may be muted in patients under general anaesthesia and the first indication of TRALI might be the appearance of copious amounts of yellow frothy sputum from the endotracheal tube.

Auscultation of the lungs will detect the presence of bilateral rales or crackles. Hypoxia, defined as PaO₂/FiO₂ <300, and the development of new bilateral lung infiltrates on chest X-ray are essential in making a diagnosis. Hypoxia may also manifest as cyanosis or oxygen saturation of <90% on room air by pulse oximetry.

New acute lung injury (ALI) is defined by the development of new bilateral lung infiltrates on the chest X-ray. The chest X-ray may show ‘white out’, a radiographic finding in which both lungs show uniform white opacities throughout. More commonly, pulmonary infiltrates are located peripherally, especially in both lower lung fields (Figure 9.1). Because some patients with TRALI have acute transient leucopenia (neutropenia) around the time of symptom onset, a complete blood count with a white cell count differential can be a useful adjunct test.

Fig 9.1 Chest X-rays of a patient with transfusion-related acute lung injury: (a) 1 day before a platelet transfusion and (b) shortly after transfusion showing diffuse bilateral shadowing of the lungs and a normal-sized heart. Reproduced with permission from AE Virchis, RK Patel, M Contreras, C Navarrete, RS Kaczmarski & R Jan-Mohamed. Lesson of the week: acute non-cardiogenic lung oedema after platelet transfusion. British Medical Journal, 1997; 314:880.

c09f001

Types of blood components that can cause TRALI

TRALI has been caused by all types of blood components, including red cell concentrates, FFP, platelet concentrates, platelets collected by apheresis, cryoprecipitate and, rarely, intravenous immunoglobulin. Plasma-rich components, namely FFP, platelets collected by apheresis and platelet concentrates collected by the buffy coat method, which are resuspended in a large amount of plasma from one of the platelet donors, pose a greater per-unit risk than plasma-poor components (e.g. red cell concentrates) [10]. Plasma that has been treated by the solvent-detergent (SD) method that is currently in use in Europe, manufactured by pooling a large number of plasma units and thus diluting the leucocyte antibodies contained in donor blood, has not been shown to cause TRALI [11]. Leukocute-reduced cellular components can cause TRALI. It is unclear if the length of storage of cellular blood components influences the risk of TRALI occurrence.

Pathogenesis

Two different mechanisms – antibody-mediated and non-antibody-mediated – have been postulated as causes of TRALI [8]. It appears that cases caused by the antibody mechanism are of greater clinical severity and more often require mechanical ventilation. In some cases, TRALI may be due to the combined effects of both mechanisms.

· Leucocyte antibodies may bind to the recipient's neutrophils, which possess the corresponding cognate antigen/s, causing them to aggregate in the pulmonary vasculature, or may bind to endothelial cells, leading to neutrophil adherence and neutrophil activation. Activated neutrophils subsequently release cytotoxic enzymes, which lead to increased vascular permeability and intra-alveolar oedema. Antigen–antibody complex formation may also lead to increased vascular permeability through complement activation. A given threshold dose of antibody might be sufficient to cause TRALI while a subthreshold dose may serve as the second ‘hit’ in a two-hit model of TRALI or cause a milder form of pulmonary injury that does not meet the definition of TRALI, a much more severe form of lung injury.

· Non-antibody-mediated TRALI results from transfusion of bioactive substances, which accumulate in cellular blood components during their storage. These bioactive substances include lipids (lysophosphatidylcholines), lipopolysaccharides, cytokines (IL-6 and IL-8), secretory phopholipase2 (sPLA2) and soluble CD40 ligand [12]. This mechanism of TRALI has been termed the two-hit hypothesis. The first hit is a patient stressor (e.g. surgery or sepsis) that causes the expression of endothelial cell adhesion molecules, leading to neutrophil adherence and activation in the pulmonary microvasculature. The second hit, consisting of passive administration of bioactive substances in stored blood components, leads to intravascular release of neutrophil enzymes, causing increased vascular permeability and resultant intra-alveolar oedema. Laboratory tests to measure bioactive substances are not widely available.

It is unclear what percentage of TRALI cases are due to each of these mechanisms. Most (but not all) series report that 80–85% of TRALI cases are caused by the antibody mechanism, with the large majority of these due to donor leucocyte antibodies directed at HLA or neutrophil-specific antigens present on the recipient's cells [13]. Some cases have been due to recipient antibody reacting with leucocytes in the transfused unit; however, this phenomenon is now very uncommon due to the widespread use of leucocyte-reduced blood components. The remaining 15–20% of cases are thought to be due to the non-antibody-mediated mechanism. Implicated antibodies include HLA class I antibodies directed against A, B and possibly C locus antigens, HLA class II antibodies directed mostly against DR antigens and neutrophil antibodies directed against human neutrophil antigens (HNAs). Involved donors can have multiple types of antibodies. The presence of a cognate antigen in the recipient that corresponds to the antibody specificity in the involved donor or a positive cross-match between the donor's serum and the recipient's leucocytes provide strong support for the serologic diagnosis of antibody-mediated TRALI. The presence of multiple antibodies in the involved donor that have the corresponding matching antigens in the recipient may be seen. It is possible that the presence of multiple donor antibodies corresponding to multiple cognate antigens poses a greater risk of TRALI than does a single congate antibody [14]. HNA-3a (formerly designated as 5b) antibodies are rare but are important to detect as these have been associated with several fatal cases of TRALI [15]. Currently, neutrophil antibody detection assays are not widely available and are just beginning to become automated. It has been reported that anti-HNA-3a may be missed unless a leucoaggultination assay or an enhanced immunofluorescense assay is used.

Lung histology

Fatal cases of TRALI show massive alveolar oedema on histological examination (Figure 9.2) [12]. Alveolar–capillary membrane disruption is widespread with hyaline membrane formation. Interstitium and alveoli are infiltrated with inflammatory cells consisting of neutrophils and macrophages. The diffuse alveolar damage resembles findings seen in ARDS from other causes.

Fig 9.2 Thin sections of fixed lung from a patient with transfusion-related acute lung injury. There is acute diffuse alveolar damage with intra-alveolar oedema and haemorrhage. There was no histological evidence of infection and all post mortem cultures (bacterial, viral and fungal) were negative. Magnification: (a) ×40; (b) ×440. Reproduced with permission from C Silliman et al. The association of biologically active lipids with the development of transfusion-related acute lung injury: a retrospective study. Transfusion, 1997; 37: 719--726.

c09f002

Differential diagnosis

Diagnosis of TRALI remains difficult because patients who experience severe respiratory distress during or after transfusion are often quite ill, have multiple other morbidities, may have cardiac or pulmonary compromise and could be suffering from conditions that are known to cause ALI or ARDS. Clinical evaluation should include investigation of other causes of ALI, which include sepsis syndrome (with or without septic shock), trauma, aspiration, smoke inhalation, near drowning, pneumonia, systemic inflammatory response syndrome, pancreatitis, postcardiopulmonary bypass and drug overdose. In any given patient, the presence of one or more of these other causes of ALI makes the diagnosis of TRALI quite difficult. A Consensus Conference has recommended that ALI occurring within 6 hours of transfusion in a patient with other ALI risk factors be designated as possible TRALI, as it is often extremely difficult to determine whether it was the transfusion or the alternate risk factor that caused the ALI [16]. The diagnosis of TRALI is difficult, if not impossible, to make in patients with pre-existing ALI.

The lack of certain clinical findings helps in the differential diagnosis between TRALI and transfusion-associated circulatory overload (TACO) [17]. This latter syndrome consists of cardiogenic pulmonary oedema, which may show one or more of the following features:

· positive fluid balance, weight gain, orthopnoea or paroxysmal nocturnal dyspnoea;

· peripheral oedema;

· hepatomegaly;

· hepatojugular reflux, heart murmur;

· new onset cardiac gallop rhythm (S3 and S4);

· an increased jugular venous pressure;

· elevated pulmonary artery wedge pressure (≥18 mm Hg) in invasively monitored patients;

· an enlarged heart with or without pleural effusion on the chest X-ray;

· radiographic appearance of pulmonary infiltrates that are more central with or without Kerley septal lines;

· echocardiograph findings of decreased systolic ejection fraction or diastolic dysfunction;

· a 50% elevation of B-type natriuretic peptide (BNP) in a posttransfusion-versus-a pretransfusion sample supports TACO whereas a BNP level of <250 pg/mL measured immediately after the onset of acute pulmonary oedema supports the diagnosis of TRALI. In critical care patients, BNP levels may be higher in patients who develop TACO compared to those who develop TRALI, yet may have a limited diagnostic value due to a large overlap among the observed values in these patient groups [18];

· some TRALI patients experience low grade fever and acute onset of leucopenia and these features, if present, suggest TRALI rather than TACO.

In addition to TACO, other conditions that can mimic TRALI include anaphylactic transfusion reactions and sepsis from transfusion of bacterially contaminated blood components. Respiratory stridor, localized or generalized skin rash, hypotension and/or shock favour a diagnosis of anaphylactic reaction. High fever, chills, rigor, shock, disseminated intravascular coagulation, a positive gram stain and culture from the transfused blood component and positive blood cultures from the recipient support a diagnosis of transfusion-transmitted bacterial sepsis. Finally, a low grade fever seen in TRALI must be differentiated from a haemolytic transfusion reaction. A clerical check of the transfusion episode showing the lack of any error, plus an absence of visual haemolysis in the posttransfusion serum or plasma and a negative direct antiglobulin test, suggest that a haemolytic transfusion reaction is unlikely.

More recently, the new term ‘transfusion-associated dyspnoea’ (TAD) has been coined for those cases of dyspnoea following transfusion that do not fit into any of the known transfusion reaction categories. A threshold model of TRALI can explain mild–moderate–severe symptoms as a consequence of the strength of the antibodies or other mediators combined with the predisposing ‘activation state’ of the patient's neutrophils and/or endothelium [8]. Further studies are needed to understand fully the entire clinical spectrum of posttransfusion lung injury.

Clinical information helpful in the differential diagnosis of TRALI is listed in Table 9.1.

Table 9.1 Clinical information that may assist in differential diagnosis of pulmonary transfusion reactions.

Clinical parameter	Interpretation
History	Underlying cardiac dysfunction and positive fluid balance may suggest TACO. Sepsis or aspiration in the previous 24 hours suggests ALI and the designation of ‘possible TRALI’. IgA deficiency may suggest allergic reaction
Physical examination	Sudden elevation of blood pressure, jugular venous distension and wheezing suggest TACO. Hypotension suggests TRALI. Stridor, wheezing and urticaria suggest allergic reaction. Fever suggests febrile reaction, sepsis/bacterial contamination
Chest X-ray	Bilateral infiltrates indicate pulmonary oedema (TACO or TRALI). Cardiomegaly (cardiothoracic ratio >0.55) and increased vascular pedicle width (>65 mm) suggest TACO
Arterial blood gas analysis, arterial oxygen saturation (pulse oximetry)	PaO₂/FIO₂ < 300 (or O₂ saturation <90% on room air) meets the Consensus Conference definition of ALI (TRALI or possible TRALI)
Haemodynamic monitoring: central venous and pulmonary artery pressures	Increase in central venous (>12–15 mm Hg) or pulmonary artery wedge pressure (>18–20 mm Hg) at the time of reaction suggest TACO
Echocardiography	Systolic (ejection fraction <45%) or diastolic dysfunction suggest TACO or other cause of cardiogenic oedema
Pulmonary oedema fluid	The ratio of pulmonary oedema albumin over plasma albumin of >0.55 suggest ALI rather than hydrostatic (TACO) oedema
Beta-natriuretic peptide (BNP)	Low values of BNP (<250 pg/mL) suggest TRALI. Increase in BNP >1.5 of pretransfusion values may suggest TACO
Leucocyte and neutrophil count before and after the implicated transfusion	Sudden, transient drop in neutrophil or leucocyte count after the transfusion suggests TRALI
Response to diuretic therapy	Rapid (minutes to hours) resolution of pulmonary oedema after diuresis may suggest TACO
Timing of the reaction in relation to transfusion and other potential risk factors	Sudden onset during or shortly after the transfusion is suggestive of a transfusion reaction rather than a pulmonary complication related to another risk factor

Management

Patient management is supportive. Virtually all patients will require some sort of oxygen support with many requiring mechanical ventilation with or without intubation. There is no evidence to support the use of corticosteroids. Fluid management may include the use of intravenous fluids to correct profound hypotension. Diuretics may be indicated if blood pressure is stable and if an element of congestive heart failure or circulatory overload is present or cannot be excluded [9, 19].

Clinicians should obtain a chest X-ray, HLA phenotype and pre- and posttransfusion BNP levels in all suspected cases of TRALI. If the investigation of the involved donors fails to show the presence of and/or risk/s for leukocyte antibody, patient testing for leucocyte antibody may be considered. Patients with TRALI resulting from their leucocyte antibodies may benefit from transfusion of leucocyte-reduced blood components. Further transfusions, if needed, do not require any other special precautions since recurrent TRALI is extremely rare. The use of platelets that are stored for less than 4 days and red cell units less than 14 days have been advocated by some authors, but there are no clinical data to support the need for such products.

Outcome and morbidity

TRALI is often quite severe and patients require mechanical ventilation for adequate oxygenation in two-thirds or more of the cases. The remaining cases require oxygen therapy by nonmechanical modalities. Pulmonary infiltrates resolve in the vast majority of patients (≥80%) in 96 hours. Slow recovery of pulmonary functions occur in 10%. Mortality is approximately 6–10%. Recurrence is extremely rare. Those who recover do not have any chronic sequelae. TRALI may be associated with a decreased long-term survival in critically ill medical patients, but this may be due to an increased fatality rate during the initial hospitalization [20].

TRALI mitigation strategies

For non-antibody-mediated TRALI, no preventive steps have been recommended or undertaken. For antibody-mediated TRALI, there is general agreement that a donor who is clearly ‘implicated’ in a case of TRALI should be deferred. An ‘implicated donor’ is defined as one who is shown to possess leucocyte antibodies that correspond to the recipient's antigen/s or a donor whose serum is reactive against the recipient's leucocytes in a cross-match test. It remains uncertain if the donor should be deferred if he or she has leucocyte antibodies but cognate antigens are not present in the recipient or if the cross-match test is negative.

In the UK, steps were taken in late 2003 to reduce transfusing FFP units collected from female donors. With this intervention, the number of TRALI cases from FFP transfusion decreased from 14 cases in 2003 to 6 cases in 2004 to 1 case in 2005. In view of these and other data, in late 2006, the AABB (formerly, the American Association of Blood Banks) recommended that plasma for transfusions be prepared from donors who are less likely to be alloimmunized [10]. Therefore, plasma units for transfusion are now increasingly prepared predominantly from male donors. This strategy has now been demonstrated to have decreased the incidence of TRALI in several different countries, including the UK, USA, Germany and the Netherlands [1,2, 9, 21]. In some European countries, SD plasma has been used as an alternative product to reduce the incidence of TRALI [11].

There are probably sufficient male donor plasma units available for blood groups O and A such that transfusion needs can be met almost exclusively by transfusion of male-only plasma. For blood groups B and AB plasma, some plasma from female donors may be necessary to meet the transfusion demand. In this regard, plasma from nulliparous female donors would be preferred to plasma from multiparous donors. Group B or AB plasma from multiparous donors can also be screened for HLA antibodies before release for transfusion. Prevention approaches for platelets require different measures as there are an insufficient number of male platelet apheresis donors to achieve a practice similar to the one described for plasma transfusion. Alternatively, deferral of females who have been pregnant at least once would still result in the loss of 40–60% of female platelet apheresis donors, and such a donor loss would likely create a critical shortage. Because of these considerations, screening blood donors, especially female donors with a history of pregnancy, for HLA antibodies and then deferring those who have them from plateletpheresis donations is a strategy that has been adopted in some jurisdictions [22]. At present, techniques for neutrophil antibody identification are cumbersome and cannot be applied for screening a large number of donors.

Key points

1. TRALI is a leading cause of death from transfusion.

2. It manifests as ARDS or noncardiogenic pulmonary oedema during or within 6 hours of transfusion.

3. Leucocyte antibodies (HLA and neutrophil-specific) and neutrophil priming agents in blood components are responsible for the syndrome.

4. Treatment is supportive, but fatality occurs in 10% of diagnosed cases.

5. Those who recover show no long-term lung injury.

6. TRALI is difficult to distinguish from transfusion-associated circulatory overload (TACO). Underlying cardiac dysfunction and positive fluid balance may suggest TACO.

7. Preventive measures include exclusion of blood donors implicated in TRALI cases and reduction of the number of transfusions of plasma-containing blood components from donors who are likely to possess leucocyte antibodies.

References

1. Fatalites reported to the FDA following blood collection and transfusion. Annual summary for fiscal year 2010. Available at: http://www.fda.gov/biologicsbloodvaccines/safetyavailability/reportaproblem/transfusiondonationfatalities/ucm254802.htm (accessed August 15, 2011).

2. Knowles S (ed.) & Cohen H, on behalf of the Serious Hazards of Transfusion (SHOT) Steering Group. The 2010 Annual SHOT Report (2011). Available at: http://www.shotuk.org/wp-content/uploads/2011/07/SHOT-2010-Report1.pdf (accessed August 15, 2011).

3. Popovsky MA & Moore SB. Diagnostic and pathogenetic considerations in transfusion-related acute lung injury. Transfusion 1985; 25: 573–577.

4. Finlay-Morreale HE, Louie C & Toy P. Computer-generated automatic alerts of respiratory distress after blood transfusion. J Am Med Inform Assoc 2008 May–June; 15(3): 383–385.

5. Toy P, Gajic O, Bacchetti P, Looney MR, Gropper MA, Hubmayr R & Clifford A. Transfusion related acute lung injury: incidence and risk factors. Blood 2012; 119: 1757--1767.

6. Khan H, Cartin-Ceba R & Gajic O. Transfusion and acute lung injury in the critically ill. In: S Kleinman & M Popovsky (eds), TRALI: Mechanisms, Management, and Prevention. Bethesda, MD: AABB Press; 2008, pp. 13–42.

7. Silverboard H, Aisiku I, Martin G et al. The role of acute blood transfusion in the development of acute respiratory distress syndrome among the critically ill: a cohort study. J Trauma 2005; 59: 717–723,

8. Bux J & Sachs UJH. The pathogenesis of transfusion-related acute lung injury (TRALI). Br J Haematol 2007; 136: 788–799.

9. Bux J. Antibody-mediated (immune) transfusion-related acute lung injury. Vox Sanguinis 2011; 100: 122–128.

10. Transfusion-related acute lung injury. Association Bulletin #05-09. Bethesda, MD: AABB; 2005. Available at: http://www.aabb.org/Content/Members-Area/.

11. Sachs UJH, Kauschat D & Bein G. White-blood cell reactive antibodies are undetectable in solvent/detergent plasma. Transfusion 2005; 45: 1628–1631.

12. Silliman CC, Paterson AJ, Dickey W et al. The association of biologically active lipids with the development of transfusion-related acute lung injury. Transfusion 1997; 37: 719–726.

13. Middelburg RA, van Stein D, Briet E, van der Bom JG. The role of donor antibodies in the pathogenesis of transfusion-related acute lung injury: a systematic review. Transfusion 2008; 48: 2167–2176.

14. Hashimoto S, Nakajima F, Kamada H, Kawamura K, Satake M, Tadokoro K & Okazaki H. Relationship of donor HLA antibody strength to the development of transfusion-related acute lung injury. Transfusion 2010; 50: 2582–2591.

15. Reil A, Keeler-Stanislawski B, Geuerney S et al. Specificities of leukocyte alloantibodies in transfusion-related acute lung injury and results of leukocyte antibody screening of blood donors. Vox Sanguinis 2008; 95: 313–317.

16. Kleinman S, Caufield T, Chan P et al. Toward an understanding of transfusion-related acute lung injury: statement of a consensus panel. Transfusion 2004; 44: 1774–1789.

17. Gajic O, Gropper MA & Hubmayr RD. Pulmonary edema after transfusion: how to differentiate transfusion-associated circulatory overload from transfusion-related acute lung injury. Crit Care Med 2006; 34: S109–S113.

18. Li G, Daniels CE, Kojicic T, Wilson GA, Winters JL, Moore SB & Gajic O. The accuracy of natriuretic peptides (brain natriuretic peptide and N-terminal pro-brain natriuretic) in the differentiation between transfusion-related acute lung injury and transfusion-related circulatory overload in the critically ill. Transfusion 2009; 49: 13–20.

19. Moore SB. Transfusion-related acute lung injury (TRALI); clinical presentation, treatment, and prognosis. Crit Care Med 2006; 34(Suppl.): S114–S117.

20. Li G, Kojicic M, Reriani MK, Fernandez Perez ER, Thakur L, Kashyap R, Van Buskirk CM & Gajic O. Long-term survival and quality of life after transfusion-associated pulmonary edema in critically ill medical patients. Chest 2010; 137: 783–789.

21. Wiersum-Osselton JC, Middelburg RA, Beckers EAM, vanTilborgh AJW, Zijliker-Jansen PY, Brand A, van der Bom JG & Schipperus MR. Male-only fresh-frozen plasma for transfusion-related acute lung injury prevention: before and after comparative cohort study. Transfusion 2011; 51: 1278–1283.

22. Kleinman S, Grossman B & Kopko P. A national survey of transfusion-related acute lung injury risk reduction policies for platelets and plasma in the United States. Transfusion 2010; 50: 1312–1321.