Risk factors, management and prevention of transfusion-related acute lung injury: a comprehensive update

Abstract

Introduction:

Despite mitigation strategies that include the exclusion of females from plasma donation or the exclusion of females with a history of pregnancy or known anti-leukocyte antibody, transfusion-related acute lung injury (TRALI) remains a leading cause of transfusion-related morbidity and mortality.

Areas covered:

The definition of TRALI is discussed and re-aligned with the new Berlin Diagnostic Criteria for the acute respiratory distress syndrome (ARDS). The risk factors associated with TRALI are summarized as are the mitigation strategies to further reduce TRALI. The emerging basic research studies that may translate to clinical therapeutics for the prevention or treatment of TRALI are discussed.

Expert opinion:

At risk patients, including the genetic factors that may predispose patients to TRALI are summarized and discussed. The re-definition of TRALI employing the Berlin Criteria for ARDS will allow for increased recognition and improved research into pathophysiology and mitigation to reduce this fatal complication of hemotherapy.

1.0. Introduction

Since the first reports describing sensitivity reactions in transfused patients in the 1950s and the association of transferred alloantibodies by transfusion in the 1980s, the understanding of the pathophysiology underlying transfusion-related acute lung injury (TRALI) has evolved [1, 2]. Most TRALI cases (80-85%), and many of those resulting in death, have been associated with the presence of anti-Human lymphocyte antigens (HLA) or anti-Human neutrophil antibodies (HNA) antibodies. All blood products, including plasma-rich (whole blood, plasma and platelets) and plasma-poor products (red blood cells RBCs, platelet concentrates, granulocytes and cryoprecipitate), have been implicated in the development of TRALI, though plasma rich products have historically been most commonly implicate [3, 4, 5, 6]. Mitigation strategies, including the practice of utilizing male-only plasma-products, plasma from females with no detectable HLA- or HNA-antibodies, and/or plasma from never pregnant females to prevent the administration of antibody-containing blood products have decreased the incidence of TRALI [6, 7, 8]. Despite these practices, TRALI remains a significant cause of transfusion-related mortality [5, 6, 7, 8, 9, 10, 11, 12]. Moreover, non-antibody-mediated TRALI is now recognized as a distinct entity and involves the recruitment and activation of neutrophils (PMNs) in susceptible (i.e. ill) patients by biologic response modifiers (BRMs) that accumulate during storage of blood products [5, 9, 12, 13, 14].

2.0. Diagnosis

The diagnosis of TRALI is solely based on its clinical presentation and depends on a high level of suspicion and vigilance at the bedside given that it is a commonly underreported entity [9]. TRALI is defined by the presence of respiratory insufficiency and hypoxemia that develop during or within 6 hours of the transfusion of blood or blood products, and imaging will reveal bilateral fluffy infiltrates consistent with pulmonary edema [9, 15, 16]. Because the 1994 American European Consensus Criteria for diagnosing ALI/ARDS was updated by pulmonary medicine experts in 2012, now the Berlin criteria, and the term ALI was dropped and replaced by ARDS, TRALI needs to take into account these new diagnostic criteria (Table 1) [17, 18]. The risk factors for ARDS included in the Berlin criteria are not oriented towards the transfusion setting and have been modified by a panel of Transfusion Medicine experts and clinicians for a consensus redefinition, please see the Definition section [19]. However, massive transfusion should not rule TRALI out and patients receiving multiple transfusions should be reviewed carefully and each transfusion event evaluated separately. To date, the diagnosis of TRALI remains as iterated in the Canadian Consensus Criteria (Table 2) and possible TRALI (p-TRALI) is still used for those cases in which a patient develops mild ARDS temporally related to a transfusion (Table 2) [20]. Further, given the decreased use of pulmonary artery, Swann Ganz, catheters, the pulmonary artery wedge pressure criterion are not often employed and the pulmonary end expiratory pressure (PEEP) >5 cm is not required for the diagnosis of TRALI. Nevertheless, because the term TRALI is firmly established in Transfusion Medicine and Hemovigilance Systems worldwide, it is unlikely it will ever be changed to Transfused-ARDS or TR-ARDS. Transfusion of blood components is common worldwide in the critically ill, and transfusion was reported in 1983 to be the most common event prior to the development of ARDS with little change since this report [21].

Table 1.

BERLIN definition for ARDS [18]

Timing	Within 1 week of a known clinical insult or new or worsening respiratory symptoms
Chest imaging	Bilateral opacities—not fully explained by effusions, lobar/lung collapse, or nodules
Origin of edema	Respiratory failure not fully explained by cardiac failure or fluid overload Need objective assessment (e.g, echocardiography) to exclude hydrostatic edema if no risk factor present
Oxygenation
Mild	200 mm Hg < PaO2/FIO2 ≤ 300 mm Hg with PEEP or CPAP ≥5 cm H2O*
Moderate	100 mm Hg < PaO2/FIO2 ≤ 200 mm Hg with PEEP ≥5 cm H2O*
Severe	PaO2/FIO2 ≤ 100 mm Hg with PEEP ≥5 cm H2O*

^*Altitude Modification = (300 mm Hg) X (Measured Barometric Pressure mmHg/760 mmHg)

PEEP Modifications: The PEEP ≥ 5 cm H₂O or CPAP ≥ 5 cm H₂O may not be required for the definition of TRALI.

Table 2:

TRALI re-definition 2019.

TRALItype I

1. Acute onset

a. Hypoxemia

PaO2/FiO2 ≤ 300* or SpO2 <90% on room air

b. Evidence of bilateral pulmonary edema (chest radiograph, CT, or ultrasound)

c. No evidence of left atrial hypertension

d. No temporal relationship with a risk factor for ARDS

2. Onset during or within 6 hours of transfusion of a blood product

3. No temporal relationship to a risk factor for ARDS

TRALI type II

1. Patients who have possible risk factors for ARDS (not major) who have not been diagnosed with ARDS or have mild ARDS: PaO2/FiO2=200-300 mmHg.

2. Clinical findings as described in 1 and 2 above.

3. Stable respiratory status for ≥12 hours preceding the transfusion

^*Altitude Modification = (300 mm Hg) X (Measured Barometric Pressure mmHg/760 mmHg)

3.0. Differential diagnosis

Respiratory distress is a common finding in multiple transfusion reactions so careful consideration of the clinical features and differential diagnosis is important. Other types of transfusion-related reactions must be ruled out including transfusion-associated circulatory overload (TACO), the leading cause of transfusion-related mortality, transfusion-associated dyspnea (TAD), transfusion-related bacterial sepsis, and severe allergic reactions.

Patients should not have evidence of fluid overload, including: cardiac failure, hypertension, positive fluid balance (liters) and/or elevated BNPs (brain natriuretic peptide: BNP <300pg/mL or NT-proBNP <2000pg/mL), distinguishing TRALI from TACO, the main differential diagnosis for TRALI; however, such BNP measurements are not reliable in the intensive care unit population in which TACO is not uncommon [22, 23, 24]. TACO is the most frequently transfusion-related respiratory complication and is most common in the intensive care unit patient population. TACO is usually the result of increased hydrostatic pressure leading to cardiogenic pulmonary edema, and fluid balance is an important distinguishing characteristic as TACO patients generally show evidence of fluid overload with elevated pre-transfusion fluid balances, which often results in hypertension [25]. Imaging will likely reveal fluffy infiltrates consistent with pulmonary edema, which may or may not be bilateral [25]. There is an immunologic component to TACO because, surprisingly, the degree of positive fluid balance demonstrates less of an association with the development of TACO than the development of circulatory overload in patients without TACO [24, 26, 27]. At a single institution the introduction of universal leukoreduction decreased the incidence of TACO [28]. Increased levels of circulating IL-10 also correlate with patients who develop TACO, though the pathophysiology of TACO is still under investigation [29].

Severe allergic transfusion reactions, including the appearance of urticaria and anaphylaxis must also be ruled out. Importantly, fever, hypotension, tachycardia and cyanosis may also occur in TRALI patients; however the wheezing and stridor associated with urticarial/allergic reactions are clinically different and do not result in pulmonary edema [3, 30, 31].

Transfusion-related sepsis and severe hemolytic reactions are often difficult to distinguish from TRALI and may present with fever and hypotension [32]. Tachypnea may be observed given the increased systemic demand for oxygen [32]. Laboratory testing including gram stain and culture of the remaining blood product in the case of suspected sepsis and DAT, haptoglobin, LDH, urinalysis and indirect bilirubin analyses when hemolytic reactions are suspected will help distinguish these reactions from TRALI [32]. Lastly, when no other cause for shortness of breath and tachypnea is found, patients may be diagnosed with TAD.

4.0. Definition

In 2004, the Canadian Consensus Conference was convened under the auspices of the Canadian Blood Service to formulate a standard definition of TRALI to foster research on TRALI epidemiology and pathophysiology to develop risk mitigation [5, 33]. Two clinical entities were defined: TRALI and possible TRALI (p-TRALI) with the latter syndrome consisting of patients with a clinical presentation identical to TRALI, which included a risk factor for ARDS [5, 33]. One year later (2005) the National Heart, Lung, and Blood Institute broadened the CCC definition by including patients whose ALI worsened with transfusion and relied upon the clinical criteria and judgment to determine if patients had TRALI or ARDS [33].

Over the past fourteen years, investigators have produced a large amount of data with regard to TRALI and p-TRALI and several groups have synthesized these data and made recommendations both for and against changes to the definitions of TRALI and p-TRALI despite these definitions not having been rigorously applied [19]. In addition, both pulmonologists and intensivists redefined ALI and ARDS in a 2012 consensus conference in Berlin [18]. Thus, to encompass the last 14 years of “TRALI data” and to harmonize a redefinition with the Berlin criteria for ALI a panel of experts was convened to redefine TRALI using the Delphi Methodology [18, 19]. TRALI is almost invariably preceded by an inflammatory first event in clinical and pre-clinical studies with some of these first hits being risk factors for ARDS [19, 34, 35, 36, 37, 38]. Therefore through analyses of these data it was decided to drop the p-TRALI diagnosis and to implement clinical criteria and judgment to diagnose TRALI or ARDS with major ARDS risk factors of sepsis, non-cardiogenic shock, and massive transfusion leading to a diagnosis of ARDS [19]. The panel further subdivided TRALI into TRALI type I, without an ARDS risk factor, and TRALI type II, with an ARDS risk factor or with pre-existing mild ARDS, which has been stable for 12 hours [19]. Despite these modifications a number of inconsistencies still exist. Firstly, much of the pre-clinical work in animals used bacterial endotoxin (LPS) as an inflammatory stimulus, the first insult in the two-event model, which many believe is a surrogate for sepsis [12, 13, 39, 40, 41, 42, 43]. However, in these models, the rodent mortality is low, approaching zero, which is expected because these animals live in an environment unsuitable for humans, which should not be equated with human sepsis [12, 13, 39, 40, 41, 42, 43]. Moreover, humans may tolerate even the highest concentrations of LPS employed in these models since LPS was used to induce high fevers to treat for tertiary syphilis [44, 45, 46]. Secondly, per the Berlin criteria, many of the TRALI and ARDS risk factors do not overlap and should be considered on an individual case basis to determine relevance (Table 3) [18, 19].

TRALI presents during or within 6 hours of transfusion of one or more plasma-containing products in a patient without pre-existing ALI, now mild ARDS, or risk factors associated with ARDS [33]. These criteria should include stable pulmonary status with P/F ratios for at least 12 hours prior to the transfusion-related event [19, 33]. The United Kingdom increased the post-transfusion period from 6 to 24 hours and found no additional cases of TRALI, validating the cogent 6 hour time frame post-transfusion [12]. The diagnostic criteria for TRALI include the acute onset of bilateral infiltrates on chest radiograph and hypoxemia defined as a PaO₂/FIO₂ ≤300 mm Hg (at sea level with lesser numbers at higher altitudes, please see the formula in Table 2), regardless of positive end-expiratory pressure level or oxygen saturation of ≤90% on room air. Other associated clinical signs and symptoms include dyspnea, tachypnea, cyanosis, tachycardia, fever, and froth in an endotracheal tube [33, 47, 48, 49]. Laboratory findings may include transient leukopenia, antibodies in the donor plasma reacting with HLA class I or class II, or PMN alloantigens and increased donor plasma concentrations of lipid mediators able to prime PMNs or activate primed PMNs adherent to the pulmonary vasculature [7, 13, 47, 50, 51, 52, 53, 54, 55, 56]. Common risk factors for TRALI identified in prospective clinical trials include acute active infection, burns, shock, coagulopathy, surgery, malignancy undergoing chemotherapy with ANC >500, pneumonia requiring IC cardiopulmonary bypass or CV surgery, and massive transfusion (Table 3) [4, 17, 33, 57].

5.0. Incidence

The incidence of TRALI varies, but it is generally felt to be an under-recognized and under-reported entity. Incidence as high as 1 in 1,333-5,000 per unit transfused has been reported in North America whereas reports from Europe vary from 1 in 29,000-270,000 per unit transfused [2, 3, 4, 7, 8, 12, 20, 54, 57].

6.0. Pathophysiology

The underlying pathophysiology has been reviewed elsewhere and so only a brief overview will be provided here [23, 54, 56, 57, 58].

6.1. Antibody-mediated TRALI

It is well established that TRALI is often associated with the passive transfer of leukocyte antibodies that include human leukocyte antigens (HLA) and human neutrophil antigens (HNA) formed following exposure to foreign antigens, most commonly during pregnancy [2, 40, 43, 56, 57, 59, 60, 61]. Strategies aimed to limit the use of blood products from parous females have decreased the incidence of TRALI. High antibody titers, mean fluorescence intensity (MFI) >1,500 on flow cytometry, and strong binding to the cognate antigen appear to be important risk factors for the development of antibody-mediated TRALI [5, 62]. The importance of antibody-specificity for its cognate antigen in the recipient has been demonstrated by look-back studies involving donors with known anti-leukocyte antibodies [40, 56, 62, 63, 64, 65]. In one case involving a recipient of a single lung-transplant, TRALI only developed in the transplanted lung that expressed the cognate antigen, but not in the native lung, which did not express the cognate antigen, which eloquently reinforced the murine model of TRALI [41, 66]. HLA class I, HLA class II and HNA antibodies have all been implicated in TRALI though the mechanisms leading to TRALI is unique for each antibody [54]. HNA-antibodies can directly bind to PMNs, whereas HLA-I antibodies likely activate PMNs indirectly via interactions with the endothelium. Data from Sachs et al has demonstrated that HLA- class II antibodies also may indirectly activate PMNs through interactions with monocytes, known to express HLA-II antigens [43]. HNA-antibodies, particularly HNA-1, −2 and −3a may be implicated in severe cases as well [3, 5, 60, 63, 67]. Earlier reports indicated that HLA-II antibodies were less common in TRALI and involved in milder disease [51, 63, 68, 69]. However, a recent commentary has asserted that antibodies to HLA class II antigens and along with antibodies to HNA antigens are the most clinically relevant, responsible for the majority of TRALI cases, especially the fatal TRALI cases [70]. Although this commentary may be accurate, modeling of this pathophysiology has been fraught with the use of isolated perfused rat lungs, the introduction of human monocytes and other leukocytes into these isolated, perfused rat lungs and the infusion of antibodies from other species [3, 42, 43, 71]. Isolated, perfused rat lungs are easily injured[13], the number of circulating monocytes is usually small (0.3-1.1% of the circulating leukocytes in healthy children and adults) in both humans and animals, and the data may be marred by obvious xenobiology through the infusion of human cells and antibodies from diverse species [3, 42, 43, 71]. However, there is an in vivo model of HLA class II antibody-induced TRALI that employed a pro-inflammatory first event, which increased the surface expression of HLA class II antigens on PMNs [40]. Lastly, antibodies are not detected in all cases of TRALI and not all antibody-containing units cause TRALI, even in recipients of blood products containing cognate-antibodies, which provides the basis for the “Two-Hit” model [18, 47, 64]. One must remember that patients who develop TRALI are not clinically well, because other than chronic transfusion programs, transfusions are usually employed in ill patients.

6.2. ‘Two-Hit’ model

Two observations: 1) that transfusion of antibody does not cause TRALI in a number of patients; and 2) that PMNs are implicated in the vast majority of TRALI in humans and animal models let to the generation of the “Two-Hit” model [18, 65, 72, 73, 74]. In 1997, a retrospective study showed that PMN-priming activity in human patients with TRALI was greater in those with underlying infection, cytokine administration, recent surgery or massive transfusion when compared with transfused-controls who did not develop TRALI [75]. Investigations in rat models demonstrated that LPS-treated rats simulating sepsis, but not saline-treated controls, developed ALI and PMN-sequestration in the lungs after exposure to the plasma from red blood cells stored for 42 days, lipid extracts of this plasma, lysophosphatidylcholines (lysoPCs) one of the lipid compounds that accumulate during routine storage [13, 39]. The addition of N-formylmethionyl-leucyl-phenylalanine (fMLF), a bacterial chemotactic protein, potentiated PMN-activation by anti-HNA-2a in an ex vivo animal studies [42]. The Two-Hit model poses that a “first hit” occurs in a patient with clinical risk factors such as recent surgery, active infection, history of chronic alcohol abuse, etc., creating a clinical milieu that effectively increases susceptibility to TRALI that develops with subsequent transfusion [4, 5, 34, 35, 37, 38, 51, 76, 77]. This clinical “first hit” promotes a pro-inflammatory environment leading to activation of the pulmonary endothelium (increased ICAM-1 expression), promotion of PMN priming and emigration to the lung vasculature with subsequent adherence of primed PMNs to activated endothelial cells [4, 13, 40, 56, 78]. The “second hit” is induced by biologic response modifiers or antibodies passively transferred via transfusion which cause the primed/adherent PMNs to release cytokines and reactive oxidative species, causing pulmonary endothelial damage, capillary leak and lung injury [4, 13, 40, 50, 56, 78].

6.2. Threshold model

The threshold model may be viewed as a variation or extension of the two-hit model and posits that mediators of TRALI, including blood product transfusion, work additively to overcome a threshold at which point PMNs become primed and activated to induce lung injury/ARDS. This model takes into account relative patient injury or predisposition as well as the strength of one or more mediators, including antibodies and may be extended to the activity of BRMs [3, 4, 58, 79].

7.0. Patient Risk factors (Table 3a)

Table 3a:

Patient risk factors associated with TRALI and ARDS (per BERLIN definition[18])

Risk Factors	TRALI risk factor	ARDS risk factor per the Berlin definition[18]
Major Risk Factors for ARDS
Sepsis	Yes[19]	Yes
Non-cardiogenic shock	Yes[19]	Yes
Massive transfusion	Yes[19]	Yes
Risk factors for TRALI
Cardiac surgery	Yes[4, 38]	No
Increased pre-transfusion plasma IL-8 levels	Yes[5]	No
Mechanical ventilation with peak airway pressure >30 cm H2O	Yes[5]	No
Chronic alcohol abuse	Yes[5, 38]	No
Current smoker	Yes[5]	No
Positive fluid balance	Yes[5]	No
Higher APACHE II score	Yes[35, 37]	No
Increased age	Yes[38]	No
End stage liver disease	Yes[34]	No
Post-partum hemorrhage	Yes[4, 80]	No
Liver transplantation surgery	Yes[5, 134]	No
Thrombotic microangiopathy	Yes[80]	No
Surgery requiring multiple transfusions	Yes[135]	No
Hematologic malignancy	Yes[4, 37, 80]	No
Other Risk factors for ARDS
Pneumonia	No	Yes
Aspiration of gastric contents	No	Yes
Inhalational injury	No	Yes
Pulmonary contusion	No	Yes
TRALI	No	Yes
Pulmonary vasculitis	No	Yes
Drowning	No	Yes
Major trauma	No	Yes
Pancreatitis	No	Yes
Severe burns	No	Yes
Drug overdose	No	Yes

The role of recipient factors is increasingly proven to be important in the pathogenesis of TRALI, particularly in high-risk patient populations such as those who had recently undergone surgery, liver transplantation, malignancy, and post-partum hemorrhage and specific clinical scenarios that include post-partum hemorrhage and autologous stem cell transplantation [4, 5, 80]. This aligns with our current understanding of TRALI as a “two hit” process whereby a pro-inflammatory condition in the patient serves as the “first hit” and the transfusion is the “second hit.” These risk factors are summarized in Table 3a and are compared with known ARDS risk factors as per the Berlin Criteria for ARDS. Elevated pre-transfusion IL-8 levels, consistent with a pro-inflammatory state, are a risk factor for TRALI [4, 5]. Other patient risk factors identified by the TRALI study group include: shock, smoking, current alcohol abuse, liver surgery, positive fluid balance, renal failure and an elevated peak airway pressure (>30 cm H₂O) in intubated patients (Table 3) [5, 19]. In a report filed by the FDA which was comprised of all reported fatalities due to transfusions from 1997 to 2002, 58 were due to TRALI, and the most common admitting patient diagnoses were cardiopulmonary disease (36%), hematological disorder (32%), diabetes and end-stage renal disease (9%) and cancer (7%) [81]. Lastly, most investigations have concluded that neutrophils are necessary for TRALI; however, high titer antibodies to HNA-A3 HLA class II antigens may activate the NADPH oxidase in the pulmonary, vascular endothelium causing injury and capillary leak, resulting in TRALI [3, 30, 41, 47, 54, 56, 57, 67]. Thus, identification of donor antibodies to the HNA-3a antigen and their relative strength must be investigated.

7.1. Critically ill patients

Multiple studies have demonstrated that TRALI occurs with a much higher frequency in the critically ill compared with general hospitalized populations [35, 53]. In studies of the critically, ill, common patient-associated risk factors included surgery, chronic alcohol use, sepsis, pneumonia, severity of illness as determined by the Acute Physiology and Chronic Health Evaluation (APACHE) III scores, aspiration events and admission to the intensive care unit (ICU) for gastrointestinal bleeding secondary to end stage liver disease [34, 51]. Reports from the Canadian Blood Services (CBS) that analyzed 305 TRALI and pTRALI cases, classified per the Canadian Consensus Conference (CCC) definitions, further described surgery as a major risk factor for TRALI comprising 38% of all TRALI cases. They further reported that cardiac surgery requiring cardiopulmonary bypass (25.0%), general surgery (18.0%), orthopedic surgery (12.5%), gynecologic surgery (9.8%) and trauma-related surgery (7.1%) were the most common sub-groups. Male gender (53.6%) was associated with more TRALI cases, though male gender may be confounded by the observation that males are more willing to undergo surgical procedures compared with females.

7.2. Major trauma

Although major trauma has been considered as a risk factor for TRALI, the association may be correlative rather than causal. Transfusion is an independent risk factor for morbidity and mortality in trauma patients [82, 83]. Post-injury ARDS occurs in 25-50% of severely injured patients (injury severity scores >15), all of whom are transfused, and occurs 24-72 hours post-injury, either alone or in a minority of these patients as part of post-injury multiple organ failure (MOF) syndrome [84, 85, 86]. Thus, despite its apparent relationship to transfusion, the overwhelming majority of ARDS occurs beyond the 6 hours post- transfusion, which is specific for TRALI and may be elicited by older stored components that induce the pro-inflammatory activation of pulmonary endothelium, an activation that is effective and durable [87].

7.3. Cardiac surgery

Several large studies have shown that cardiac surgery is a high-risk clinical scenario for the development of TRALI and cardiopulmonary disease is associated with a large percentage of TRALI-related mortality [4, 37, 38, 81]. Intrathoracic surgery requires that the lungs are deflated for a period of time, often for several hours, after which time the lungs are reflated,[38] likely inducing localized injury to the pulmonary vascular or parenchyma. Furthermore, cardiovascular bypass has been shown to cause PMN priming [88].

7.4. Elderly patients

Elderly patients appear to be at a higher risk for the development of TRALI. ‘Inflamm-ageing’ is a relatively new concept referring to chronic low-grade inflammation associated with higher baseline levels of pro-inflammatory cytokines that include IL-8, which is a risk factor for TRALI, described in elderly people [89]. A retrospective analysis of Medicare data found that persons greater than 79 years of age actually had a lower odds of TRALI compared with persons ages 65 to 79 years [90]. This finding may reflect a waning innate PMN dysfunction associated with ageing [89].

7.5. Genetic factors

Some evidence exists that genetic mutations, frequently those associated with inflammatory proteins such as angiotensin-converting enzyme (ACE), mitogen-activated protein kinase kinase kinase 1 (MAP3K1), surfactant protein B (Sp-B), interleukin-6 (IL-6), interleukin-10 (IL-10), the cation channel transient receptor potential vanilloid (TRPV) 4, liprin alpha (PPFIA1) and tumor necrosis factor-alpha (TNF-α) may be associated with predisposition to or protection against ALI/ARDS, but no specific associations with TRALI have been reported. More extensive studies need to be conducted to make any conclusions regarding genetic contribution in the development of TRALI [91, 92, 93, 94, 95, 96, 97].

7.6. Emerging/possible patient risk factors

New human patient and animal model data is emerging that may implicate low levels of interleukin-10 (IL-10) as a risk factor for the development of TRALI. One study found that IL-10 levels were lower in TRALI patients as compared with patients who developed transfusion associated circulatory overload (TACO) [29]. A small observational study showed that IL-10 levels were lower in transfused patients with and without TRALI compared with non-transfused septic patients with ALI [98]. Interactions with the gastrointestinal (GI) microbiome are under exploration and may further enhance our understanding inherent patient susceptibility to TRALI [99].

8.0. Transfusion-related risk factors (Table 3b)

Table 3b:

Transfusion-risk factors associated with TRALI

High volume of female plasma[5]

Cognate HLA Class II antibody[5]

Cognate HNA antibody[78]

Positive HNA antibody by GIFT[5]

Cognate HLA Class I antibody[63]

8.1. Female donors

Receipt of plasma or whole blood from female donors are known risk factors for TRALI. The incidence of TRALI has decreased with the implementation of national policies designed to exclude females from donating for high plasma-volume products in the UK, the Netherlands and US [5, 12, 100].

8.2. HNA and HLA Class-II antibodies

The report from the TRALI Working Group determined that elevated volume of HLA class II antibody and the volume of cognate anti-HNA antibody (positive by granulocyte immunofluorescence test) were both positively associated with TRALI [5]. The effect of transfusion number, a previously reported risk factor, was attenuated in multivariate in which the effects of HNA- and HLA-II antibodies predominated, in addition to receipt of plasma or blood from a female donor. Older age of blood, non-cognate or weak cognate class II antibody or class I antibody were not found to be a positive risk factors in multivariate analysis.

9.0. Prevention and mitigation

Once, the leading cause of transfusion associated mortality, the incidence of TRALI has been significantly reduced with the implementation of leukoreduction technology and TRALI mitigation strategies that include the deferral of high risk donors (females with a history of pregnancy), a male-only plasma donor population, and screening for antibodies implicated in TRALI with the subsequent deferral of these donors [5, 7, 100, 101, 102]. The FDA reported a decrease in TRALI cases from 35 reported in 2006 to 8 cases reported in 2016 [10]. Likewise, the most recent SHOT data shows a decrease in TRALI from 24 cases reported in 2003 to 3 cases reported in 2017 [12]. Despite this, TRALI remains a significant cause of transfusion-related morbidity and mortality despite various reports demonstrating an almost 4-fold drop in TRALI incidence over the past several years [5, 101]. As our understanding of TRALI evolves, mitigation strategies will need to as well.

9.1. Leukoreduction

Results of the TRAP (the Trial to Reduce Alloimmunization to Platelets) study, published in 1997, provided strong evidence for the reduction of alloimmunization and platelet refractoriness by utilizing leukoreduction by means of white blood cell filtration [103]. Proving so successful, universal leukocyte reduction was introduced in 2000, after which time reports of TRALI significantly decreased. One retrospective analysis compared complications from the pre-leukocyte reduction era to the post-leukocyte reduction era and found an 83% reduction in TRALI reporting (2.8 to 0.48 events per 100,000 blood products transfused) after the introduction of universal leukocyte reduction [28]. Leukoreduction of blood components decrease CMV transmission, HLA antibody formation and febrile non-hemolytic transfusion reactions [104]. A reduction in the accumulation of multiple pro-inflammatory mediators, such as IL-6, IL-8, MCP-1, sCD40 ligand, and biologically active lipids, which have been implicated in the pathogenesis of TRALI are all reduced by leukoreduction [104, 105, 106, 107, 108, 109].

9.2. Female donor exclusion and exclusion of other “high-risk” donors

As anti-HLA and anti-HNA antibodies have been implicated in the majority of TRALI cases, including most lethal cases, many countries began to institute donation policies aimed to prevent donations from “high-risk” donors. In 2003, the National Blood Service (NBS) in the United Kingdom began using male donors for FFP and buffy coat-derived platelet pools.[100] Retrospective analyses of cases reported to the Serious Hazard of Transfusion (SHOT) scheme demonstrated a dramatic drop in TRALI from 15.5 to 3.2 per million FFP units and from 14.0 to 4.8 per million pooled platelets units after the introduction of these strategies [7]. After the adoption of male-only plasma donation in the Netherlands and Switzerland in 2007, cases of reported TRALI decreased by 33% and 24%, respectively [102]. In 2007, the American Red Cross adopted a male-predominant plasma donor strategy and subsequently observed an 80% reduction in TRALI associated with plasma transfusion [110]. The authors note that due to a continued reliance on female donors to meet the demands for AB plasma, TRALI association with this product has not been reduced. Reflective of these findings, in 2014 the AABB further tightened plasma restrictions, mandating plasma only donated by males, never-pregnant females or confirmed HLA antibody negative women who have been pregnant. On October 1^st of 2016, the same rules were extended by the AABB to include apheresis platelets.

9.3. Screening for and exclusion of HLA- and/or HNA-antibody containing units

The exclusion of all female donors may unnecessarily exclude a fraction of ‘safe’ donors [57]. Many blood banks allow females without a history of pregnancy to donate and allow ever-pregnant females to undergo anti-leukocyte antibody testing. However, donor screening studies show that 1.7-7.8% of never-pregnant females and up to 7.1% of never-transfused males will test positive for antibodies [63, 74, 100]. A look-back study in Germany showed that HLA class II, HNA and HLA-A2 antibodies were responsible for the majority of clinically relevant TRALI cases [63]. Unlikely to be feasible, the most conservative approach may be to test all donors for the presence of antibodies.

9.4. Restrictive use of blood products and blood product management

Although transfusion is critical in many clinical situations, its administration continues to be associated with adverse reactions that include TRALI [5, 9, 12, 25, 32, 52, 53]. With respect to ARDS, a restrictive approach to transfusion in the critically ill has been shown to be a protective measure and this may translate to TRALI as transfusion is the causative “second hit” [111].

9.5. Platelet additive solutions (PAS)

Platelets are stored in gas-permeable bags that allow for the release of CO₂, a byproduct of free fatty acid metabolism and glycolytic production of lactic acid accumulated during storage, helping to maintain a pH above 6.2 [32]. An unforeseen benefit of platelet plasma reduction has been the concurrent reduction in HLA- and HNA-antibodies [32]. LysoPCs have been shown to accumulate during storage of platelets in plasma compared to fresh plasma [38]. A study conducted in the Netherlands suggests that when buffy-coat-derived platelets are stored in PAS, the incidence of TRALI is lower than what has been traditionally observed from buffy-coat derived platelet transfusion, but is similar to what is observed in apheresis-derived platelets. More work is needed to establish the role of PAS in reducing TRALI.

9.6. Washing

Washing may remove excess potassium, cell-free hemoglobin and RBC microparticles and additionally, may help to remove RBCs susceptible to osmotic fragility and hemolysis. Washing may also reduce cytokines and complement proteins [112]. In vitro data performed in HUVECs (human umbilical vein endothelial cells) showed that washed RBC exposure was associated with decreased endothelial surface expression of CD62E (E-selectin) and CD106 (VCAM) suggesting yet another benefit to washing with regard to TRALI prevention [112]. A single institution, retrospective study demonstrated that washing resulted in a complete absence of TRALI cases in more than 28,000 transfused units [28].

9.7. Pathogen reduction technology

Pathogen reduction technologies (PRTs) have improved the safety of blood products over the past several decades. These technologies generally work by targeting pathogen nucleic acid or lipids within membranes [113]. Known pathogens transmissible by allogeneic blood product transfusion as well as not yet unidentified and potentially newly emerging pathogens are all inactivated, providing a major advantage to the use of PRT in the blood bank. The Intercept™ system was developed by Cerus in 2002 and inactivates viruses, bacteria, protozoa and leukocytes by utilizing an amotosalen and UVA [114]. Findings from a large hemovigilance program involving multiple centers in 11 countries demonstrated that Intercept™-platelets are both safe and effective with an overall low rate of adverse transfusion reactions [114]. No TRALI cases were associated with 19,175 Intercept™-platelet transfusions in 4,067 patients [114]. The Riboflavin- and UV light treatment (Mirasol Pathogen Reduction Technology®) system also appears to be a safe and effective option for bacterial and viral pathogen inactivation [115]. RBCs treated with this technology and stored to >25 day, have less hemolysis, free hemoglobin, and ADP levels compared with control RBCs as well as reduced antibody formation to platelet antigens [116, 117]. A recent murine study showed that the use of aged Mirasol PRT-treated RBCs failed to significantly increase lung injury in an antibody-mediated model when compared to standard-issue RBCs stored to the same age [118]. Collectively these data show that PRT is safe and effective in preventing pathogen transmission and may decrease the risk of TRALI in blood product recipients.

9.8. Solvent/detergent (S/D) plasma

Solvent/detergent (S/D) plasma (Octaplas®) is comprised of plasma pools from 500-1600 donors that undergoes pathogen inactivation followed by filtration. Due to a dilutional effect, leukocyte antibody concentrations are likely reduced to clinically negligible levels. Indeed, since the introduction of this product in Norway in 1993, no cases of TRALI were reported over a 10-year period [119]. A consensus panel of experts in Italy recently reviewed the existing literature regarding the efficacy and safety of using solvent detergent plasma (SD-plasma). Therefore, the use of SD-plasma, where available, may reduce the incidence of TRALI.

9.9. Prevention strategies in development

Standard leukoreduction filters are expected to reduce leukocytes in RBC and apheresis platelet units to <5.0 x10⁶ cells per unit, which translates to a 3 log reduction of leukocytes and a 2 log reduction of platelets [32]. Standard filtration also reduces HLA antigen exposure, cytokine accumulation, CMV transmission and sCD40L accumulation in stored RBCs [14, 105]. An experimental filtration paper for pre-storage leukoreduction (LR), containing standard LR material and a proprietary material designed to remove antibodies and immunoglobulins in both small volume and standard volumes, was tested for its ability to reduce the proinflammatory activity of stored RBCs [120]. Blood was donated from 31 multiparous females donors with known HLA class I and/or class II antigens and 16 control donors negative for antibodies. They found that the experimental filter successfully removed >96% of IgG, 93% of anti-HLA I antigen and 99% of anti-HLA II antigen antibodies [120]. The supernatant from RBCs that underwent experimental LR resulted in reduced PMN priming activity, specifically lipid priming, compared to the standard LR model [120]. The supernatant that underwent LR with the experimental filter also caused less TRALI in a rat-model when compared with non-LR control RBC supernatant [120].

10.0. Management

10.1. Supportive care

Supportive care remains that mainstay in treating TRALI cases. This includes supplemental oxygen, non-invasive ventilator support and ventilator support as needed. Some patients have required extra-corporeal membrane oxygenation (ECMO). The reduction of ALI risk factors such as targeting lower peak pressures in ventilated patients to reduce barotrauma has shown to be effective. Unlike in cases of fluid overload associated with transfusion, diuretics have not shown to be helpful and may in fact further harm the patient by causing further hypotension. Therefore, it is recommended that fluid balance is carefully monitored and patients remain euvolemic or slightly “fluid depleted”.

10.2. Corticosteroids

Although steroids may be effective in decreasing mortality and time to achieving unassisted breathing in ARDS and ICU-free days, there is limited data demonstrating efficacy in TRALI.[121] Several ARDS studies that demonstrate steroid efficacy also reported decreases in markers of systemic inflammation [121]. However, even in ARDS patients, the role of steroids is still controversial and several large studies have not demonstrated improved clinical outcomes associated with steroid administration [121]. Animal studies suggest that methylprednisone may decrease IL-6 levels (suggestive of systemic inflammation), but not pulmonary edema, bronchoalveolar lavage fluid or protein levels [122]. Steroids have been used anecdotally in TRALI case-reports, but no large-scale studies exist [123, 124].

10.3. Immunologic work-up of involved donors and recipients

Antibody-mediated TRALI requires identification of alloimmunized donors and recipients [125]. Donor workup for the presence of anti-HLA I, anti-HLA II, and anti-HNA antibodies should be undertaken so that these donors are deferred from further donation of plasma rich products [20, 32]. As described by the ISBT Working Party on granulocyte Immunobiology, tests for antibodies against HLA class I antigens include: a) enzyme immunoassay, b) flow cytometry with microbeads, c) lymphocytotoxicity, d) lymphocyte immunofluorescence or e) other validated tests; while those assays for antibodies directed against HLA class II antigens rely on only a), b), or e) [125]. Importantly most of the TRALI reactions will be due to donor-derived antibodies to HLA class II antigens not HLA class I which may be absorbed by recipient platelets [126]. The recommendations for HNA antibody detection rely upon the combination of granulocyte immunofluorescence test (GIFT) and the granulocyte agglutination test (GAT) [125]. Ideally, testing for cognate antigen(s) in the recipient should be performed as well [20, 32]. For more information please see the ISBT Working Party consensus paper on this topic [125].

10.4. IL-10 therapy

Recent work by Kumar and colleagues indicate that T regulatory cells and IL-10 may be protective against the development of antibody-mediated TRALI suggesting a possible role for IL-10 as a therapeutic agent [127]. They were able to rescue and prevent mice with antibody-mediated TRALI by injecting IL-10 following onset of symptoms and prior to the onset of symptoms respectively [127]. The role of IL-10 in the prevention or treatment of TRALI has not been established in humans.

10.5. Anti-platelet agents

Murine models and studies in human patients are inconclusive as to the role of anti-platelet agents in preventing TRALI. In a murine TRALI-model, pre-treatment with aspirin prevented TRALI seemingly by decreasing platelet sequestration in the lungs.[41] A case-control study in critically ill adult patients failed to show a protective effect of anti-platelet agents on TRALI [128]. To date, there is not a role for anti-platelet agents for the treatment of TRALI.

10.6. Anti-complement agents

Circulating HNA-1a IgM antibodies were identified in a donor implicated in a TRALI case, that were able to bind C3d via granulocyte immunofluorescence testing (GIFT) [129]. Experimental murine models indicate that complement activation (C3a and C5a generation) as well as bronchoalveolar lavage fluid protein levels and various chemokine levels (including interleukin-6 (IL-6)) increases with the onset of TRALI [130]. However, the role of complement in TRALI is not well understood. Experimental murine models have also shown that complement may be activated in some but not all antibody mediated TRALI [42, 131].

11.0. Expert opinion

Revisions to the nomenclature as suggested by the TRALI consensus group will potentially improve congruence in TRALI recognition and adjudication; furthermore, this may improve uniformity in research studies [19]. This re-definition in turn may provide increased insight into the nuanced pathophysiology of TRALI and lead to better therapeutic interventions and preventative measures. As TRALI continues to represent an under-recognized entity in the medical community, we encourage Transfusion Medicine Providers to continue to educate our clinical colleagues on best practices in blood utilization management and recognition of adverse transfusion-reactions. Vlaar et al reported that of 16 cases of pTRALI identified in their prospective study, only one case was reported to the blood bank [38].

The introduction of low-risk TRALI donor strategies in the collection of plasma-containing products has led to a decreased incidence of TRALI. A meta-analysis by Muller et alanalyzed the results from 10 observation studies, utilizing a random-effects model, and showed that low-risk donor strategies including the use of male-only plasma, screening for absence of HLA and/or HNA antibodies and plasma from females without a history of pregnancy were protective [6]. Sub-analysis showed that this risk reduction was greatest in high-risk populations (i.e. critically ill patients and those undergoing high-risk surgeries such as cardiac surgery) providing further support for the “Two-hit” and/or multifactorial models. These strategies have not completely eliminated the incidence of TRALI, which remains an important cause of transfusion-related morbidity and mortality. Focus on product modifications as well as more advanced, high-throughput and less-expensive means of donor testing may further reduce the incidence of TRALI.

Careful treatment of high-risk patients may help to further decrease new TRALI events from occurring. Patients undergoing cardiac surgery and expected to remain on cardiopulmonary bypass for extended periods of time should receive blood products known to not contain anti-leukocyte antibodies, especially antibodies directed against HLA class II and HNA antigens [38, 70]. Furthermore, product washing and new leukoreduction technologies, particularly in high-risk patient populations may prove to benefit patients at high risk for TRALI.

11.1. Five-year view

Current RCTs underway, the WAR-PRC (the Washing of Allogeneic Red blood cells for the Prevention of Respiratory Complications) and REDWASH (Red Cell Washing for the attenuation of transfusion-associated organ injury in cardiac surgery), may help us understand whether product washing will improve outcomes in high-risk patients such as those undergoing cardiac surgery and potentially provide guidance for the management of other high-risk patients [132, 133].

As the pathogenesis of TRALI is still not completely understood, further studies will help to enhance our understanding and drive the development of new therapies and preventative strategies. Newer filter designs for leukoreduction show promise in reducing TRALI based on in vitro and in vivo rat studies that demonstrate a reduction in PMN priming and TRALI induction [120]. Novel therapies such as the use of IL-10 in patients and anti-complement agents are under active investigation [23, 57].

Article highlights.

• Diagnostic criteria for TRALI, especially the nomenclatures possible TRALI (pTRALI) is an area of contention

• New nomenclature for TRALI will likely be reported soon

• Mitigation strategies including use of “low-risk” plasma products have reduced TRALI incidence, but TRALI remains the leading cause of transfusion related mortality

• High risk patients may benefit from product modifications such as washing; clinical trials are currently underway to address this concept

• Additional clinical studies are needed to better understand pathogenesis of and identify clinical markers for TRALI