Transfusion‐associated circulatory overload and transfusion‐related acute lung injury: diagnosis, pathophysiology, management and prevention

Pulmonary oedema may be defined as the clinical manifestation of fluid accumulation in the lungs, resulting from an imbalance between fluid filtration and resorption. Pulmonary oedema leads to impaired gas exchange and may cause respiratory failure. It is considered as ‘cardiogenic’ in cases where the heart fails to remove efficiently fluid from the lung circulation, or as ‘non-cardiogenic’ in case of direct injury to the lung parenchyma [1]. Transfusion-related acute lung injury (TRALI) and transfusion-associated circulatory overload (TACO) are the two major complications of blood transfusion involving clinical signs of pulmonary oedema. Transfusion-associated circulatory overload had been recognized since the early times of transfusion [2]. In contrast, TRALI was identified much later, and although a few case reports were published as early as 1957 [3] and in the 1970s [4], the first comprehensive description of this adverse reaction was published in 1985 [5]. Differential diagnosis of transfusion-related pulmonary adverse reaction may include not only TACO and TRALI, but also anaphylactic reactions, and even bacteria contamination and haemolytic transfusion reaction may involve pulmonary manifestations at some point [6]. However, the major difficulty remains to differentiate TRALI from TACO. This differential diagnosis is essential in clinical practice, because the therapeutic approach of these two syndromes and the investigations in blood donors in order to confirm the diagnosis are totally different. Despite these two concerns, it is still frequent that post-transfusion pulmonary oedema is not adequately categorized. The extremely variable incidences described according to dedicated studies and haemovigilance results strongly suggests that up to now, TRALI and TACO remain both under recognized, and deserve active information of clinicians and transfusionists. This review intends to provide a description of these two adverse reactions related to blood transfusion, focusing on the following: clinical description and differential diagnosis; pathophysiology; incidence according to the different blood components; investigations necessary to confirm suspected cases; and means for prevention. Transfusion-associated circulatory overload may occur during transfusion or within 1 to 2 h after its completion, although it may begin lately. There is at present no consensus for a precise time limit to consider, which is not so surprising, as this adverse event is extremely dependent on patient's clinical status, independent of blood transfusion. Major clinical signs are dyspnoea, orthopnoea, tachypnoea and cyanosis associated with tachycardia and systolic hypertension. Rales are observed at auscultation and may be prominent initially in the lung bases, although they are present in all the lung fields once the pulmonary oedema is established. Chest radiograph shows interstitial infiltrate, and may reveal cardiomegaly, although not in all patients with heart failure [7]. Transfusion-associated circulatory overload management consists in stopping the transfusion as soon as the symptoms appear, and treating symptomatically by providing oxygen, using intravenous diuretics in order to obtain a rapid reduction of plasma volume. Although many cases can be managed outside an intensive care unit, some cases deserve a close monitoring for 1 or 2 days. As many patients will require to be transfused again in the future, a great attention should be given to avoid recurrence of TACO. Careful control of the patient's fluid balance previous to transfusion, prescription of no more than one red cell concentrate for a given day (in most medical patients), use of the recommended blood flow (2 ml/mn) and preventive diuretics injection in case of positive fluid balance are the major points to consider. Transfusion-related acute lung injury had been defined differently in three publications [5,8,9], the most internationally accepted one being the result of an international consensus conference organized in Toronto in 2004 [9]. The 2004 consensus conference definition of TRALI uses the The American-European Consensus Conference definition of ALI [10]. ALI is defined as acute onset of bilateral infiltrates on chest radiograph and pulmonary artery occlusion pressure (PAOP) < 18 mmHg obtained by pulmonary artery catheterization (if unavailable, lack of clinical evidence of left ventricular failure), and PaO2 : FiO2 < 300 mmHg. Transfusion-related acute lung injury is defined as ‘a new episode of ALI that occurs during or within 6 h of a completed transfusion, which is not temporally related to a competing aetiology for ALI. The diagnosis of TRALI is a clinical and radiographic diagnosis and is not dependent on the results of laboratory tests or any proposed pathophysiologic mechanisms. TRALI should currently be considered a clinical syndrome rather than a disease with a single aetiology’. The symptoms of TRALI include dyspnoea, cyanosis and tachycardia. After a few hours, the oedema is usually bilateral involving the whole lungs, as shown by chest radiograph. Fever (1 to 2°C increase) may be present, as well as moderate hypotension. However, in approximately 15% of cases, hypertension may be observed [5]. Another factor is that TRALI may be associated with a transient leucopenia, for usually less than 24 h [11]. Once TRALI is suspected, it is important to manage the patient according to the severity of this adverse event. If all patients require oxygen supplementation, not all deserve a transfer to an intensive care unit. Most cases improve after 2 to 4 days, but 20 to 30% of the patients require mechanical ventilation and a full intensive care monitoring. Although some factors such as a history of heart failure, a positive fluid balance and the absence of risk factor for developing an ALI make a clear orientation in favour of TACO, much more complex situations may occur where the classification is less easy. Pulmonary artery occlusion pressure has been considered as a key criterion in distinguishing cardiogenic from non-cardiogenic pulmonary oedema, values of 18 mmHg or below being in favour of a non-cardiogenic origin. However, nowadays, PAOP monitoring is far from being always performed. This invasive technique has been considered as associated with increased morbidity and mortality [12], and its utility in monitoring ALI is a matter of debate. Moreover, we should have in mind that each measurement may not be representative of the overall clinical situation, and that high values have been observed in situations considered as authentic ALI syndromes [13]. The fluid protein concentration analysis could in theory distinguish circulatory overload from ALI, the lower the protein concentration ratio of the alveolar fluid/plasma is, the more probable is the circulatory overload. A ratio < 0·65 strongly suggests a circulatory overload [14,15]. However, the technique is feasible only in intubated patients, and is subjected to technical difficulties that may alter the results [16]. Standard chest radiography may not be informative to distinguish TRALI and TACO. However, the measure of the vascular pedicle width and cardiothoracic ratio may help: it has been shown that a vascular pedicle width above 70 mm along with a cardiothoracic ratio above 0·55 may predict a PAOP above 18 mmHg [17]. B-type natriuretic peptide (BNP) is a cardiac neurohormone produced from the ventricules in case of volume expansion and pressure overload. Its dosage in plasma may be of some help: high BNP levels above 1200 pg/ml are clearly associated with circulatory overload with a specificity greater than 90%, and levels below 200 pg/ml associated with non-cardiogenic pulmonary oedema with a specificity greater than 90%[18]. In a case–control study, BNP has been shown a useful marker in confirming TACO diagnosis [19]: the post- to pre-transfusion BNP ratio was significantly high in patients with transfusion-associated volume overload as compared to matched recipients of blood transfusion without adverse reaction. The handicap of such an approach is mainly the fact that it requires to test pre- and post-transfusion samples, which means that the pre-transfusion sample may have been stored up to 2 or 3 days in poorly controlled conditions, potentially leading to false-positive results. This difficulty may be overcome by dosing the N-terminal fragment of the pro brain natriuretic peptide (NT-proBNP), a biologically inactive fragment cleaved from the proBNP along with the biologically active BNP. This marker has been shown to be more accurate than BNP to identify asymptomatic patients at risk for the development of heart failure [20]. Using this marker enables to work only on post-transfusion plasma. In a case–control study of 16 clinically defined TACO and 24 controls, post-transfusion NT-proBNP above 1000 pg/ml was found to be a good marker of TACO, with a sensitivity > 93% and a specificity > 83%[21]. Finally, it has to be stressed that, although found normal in a TRALI case report [22], this test has not actually been investigated in clinically defined TRALI at a large scale, which will be mandatory to define its definitive value for differentiating TRALI from TACO. Investigation of fluid balance is part of basic patient's care, and a clearly identified positive balance related to an intake excess is in favour of TACO. Of note, it has been indirectly shown by plasma NT-proBNP dosage that most patients with clinical TACO had a high NT-proBNP concentration, reflecting a positive fluid balance previous to the transfusion that caused TACO [21]. As already mentioned, some additional symptoms, such as the observation of fever, transient leucopenia, may add probability for the diagnosis of TRALI. Table 1 summarizes the factors contributing to differential diagnosis between TRALI and TACO. In TACO, pulmonary oedema is the result of a heart failure caused by increased hydrostatic pressure in lung due to an increased pulmonary blood volume, an increased central venous pressure. In transfusion practice, many patients are at risk for TACO. Infants are at high risk, as they may be subjected to high positive fluid balance. Elderly patients who may have a limited capacity to respond to a positive fluid balance, are also at high risk, as are patients with a history of heart failure. Finally, profoundly anaemic patients who have already tachycardia in order to compensate their anaemia, have a limited ability to adapt to blood volume variation, and are also at serious risk of TACO if not managed preventively. In the past 10 years, our knowledge in regard to the pathophysiology of TRALI increased tremendously, and it is now possible to propose a unified vision of it [23]. Neutrophils are the key trigger for TRALI, as the basic alveolar lesion leading to pulmonary oedema is the result of the release of preformed granular enzyme, and of the synthesis of highly toxic reactive oxygen species following neutrophil activation. Circulating neutrophils are required to be primed before activation. Priming is a process that potentiates the effects of an activating stimulus [24]. Therefore, it can be assumed that TRALI requires priming and activation of neutrophils to occur, and that at least priming or activation is provoked by the blood component transfused. A great number of agents have been described as being able to prime neutrophils. Among them, we can mention platelet activating factor, tumour necrosis factor-α (TNF-α), interleukin 8, granulocyte/macrophage-colony stimulating factor and interferon-γ. Most of these molecules may be released either by dying cells or by activated endothelial cells, monocytes and lymphocytes. In addition, numerous infectious agents, and bacteria-derived lipopolysaccharides may also act directly as priming agents. Transfused patients may be in clinical conditions where activation of endothelial cells, monocytes and lymphocytes may occur, such as recent surgery, cardiovascular disease and leukaemia. Moreover, they can have an active infection leading to priming of neutrophils too. Activation of pulmonary endothelial cells may be a prominent mechanism for neutrophil priming. Activated pulmonary endothelial cells express surface membrane receptors for L-selectin, P-selectin and ICAM-1 that favour neutrophil adhesion, and produce cytokines such as platelet activating factor and interleukin 8 which act as priming agents. A good clinical example of this activation of pulmonary endothelial cells as a starting event of TRALI is provided by the observation of a recipient of lung transplant who develop TRALI on the transplanted lung, after transfusion of a red cell concentrate containing an anti-HLA class I antibody recognizing an antigen present in the donor lung, but absent in the recipient [25]. Neutrophils can also be primed by a factor brought by the transfused blood component, which can be an antibody, a cytokine or bioactive lipids (see further). Whatever the way they had been primed, neutrophils exhibit a shape change, and their ability to cross the lung circulation is impaired. This results in an accumulation of primed neutrophils in the lung that can be further activated. Eventually, in case there is no further stimulus, neutrophils will go back to their initial state in at least 24 h [26]. Primed neutrophils may be activated by an exogenous factor brought by the transfused blood component, which can be an antibody, a cytokine or bioactive lipids (see further). Alternatively, neutrophils may be activated by already activated pulmonary endothelial cells. Actually, it is the interaction between neutrophils and endothelial cells that results in TRALI. Anti granulocytes antibodies. Antibodies against neutrophil antigens human neutrophil antigen (HNA)-1a, HNA-1b, HNA-1c, HNA-2a and HNA-3a have been observed in many TRALI cases [27]. An interesting observation that corroborates the ‘priming plus activation’ hypothesis is that not all the situations of incompatibility (i.e. where the recipient bears the corresponding antigen) shall lead to a clinical manifestation of TRALI. In the first description of this phenomenon, donations of a frequent plasma donor with anti-HNA-3a caused TRALI in 36% of the recipients, while the corresponding antigen is present in > 90% of Caucasians [28]. In a more recent observation, out of 10 patients receiving platelets from a donor with anti-HNA-3a, TRALI was observed on two occasions, but in a single patient [29]. Another interesting observation had been done in an ex vivo model, using an anti-HNA-2a antibody. HNA-2a expression in man is heterogeneous, some individuals having an expression on 70% or more of the neutrophils, while others express it on less than 30%. Using neutrophils from ‘highly expressed’ HNA-2a individuals, it was possible to induce neutrophil activation directly by adding the antibody, while using neutrophils from ‘weakly expressed’ HNA-2a individuals, neutrophil activation required the addition of a known priming agent to occur [30]. Anti-HLA class I antibodies. These antibodies not only are able to react with neutrophils, but also with endothelial cells. Moreover, it is clear that at least some of them are able to prime and to activate neutrophils. Indeed, anti-HLA class I antibodies are commonly found in look-back studies of TRALI cases. As these antibodies are frequently found in blood donors, it is important to test that the corresponding antigen is present in the recipient in order to confirm their immutability. Anti-HLA class II antibodies. First described as related to TRALI in 2001 [31], there is now evidence that these antibodies may play a role in the onset of TRALI. However, it is unlikely that they act directly either on neutrophils or on endothelial cells, as the corresponding antigens are not expressed on these cells. The hypothesis that these antibodies could act indirectly by first activating monocytes, and that activated monocytes could provoke the priming of neutrophils, although elegant, is still not fully documented. Bioactive lipids. The hypothesis that breakdown products membrane lipids such as phosphatidylcholine species could activate neutrophils through their lyso-platelet activating factor receptor was proposed in 1994 [32]. Additional argument in favour of such a role was provided by an in vitro model [33]. In addition, the description of a TRALI case in a recipient of a pre-deposit autologous red cell concentrate showed clearly that TRALI could happen in the absence of immunologic incompatibility [34]. Cytokines. Apart from the already mentioned cytokines, CD40 ligand (CD40L) is a platelet-derived mediator present in platelet concentrates as a soluble form (sCD40L). As CD40 is expressed not only by monocytes and macrophages, but also by neutrophils, it has been hypothesized that it may act as an activation means for neutrophils leading potentially to TRALI [35]. Neutrophils. In the modern blood transfusion, using leucoreduced blood component, the amount of neutrophils in blood components is negligible. However, some patients receive specific transfusions of neutrophils, and in theses cases, TRALI had been observed, as clearly related to an immunological incompatibility between the transfused neutrophils and the presence of a corresponding antibody in the recipient [36]. Mysterious 15 years ago, the underlying mechanisms for TRALI are much more understood nowadays. Priming and activation of neutrophils is the key event produced by a great variety of inducers, at least one being present in the transfused blood component responsible for TRALI. It is clear from several studies targeted to detect TACO that it is a frequent adverse reaction after transfusion of blood components. In a first study by Popovsky, 1% of patient undergoing surgery developed TACO, which means approximately 1300 per million of red cell concentrate transfused [37]. In a more recent study in intensive care unit, TACO incidence was estimated at 2800 per million of red cell concentrate transfused [38]. Data from haemovigilance networks of France and Québec provide also useful information. In France [39], a mean of 192 TACO have been reported annually between 2000 and 2006 (extremes: 168–209), which corresponds to an incidence of 76 per million blood components transfused (92% of TACO were related to red cell concentrate transfusion). Among these cases, the rate of death is 2%, which represents three to four TACO-related deaths per year. In Québec [40], 159 TACO have been reported in the 2 years 2003–2004, which corresponds to an incidence of 242 per million blood components transfused. In conclusion, there is no doubt that TACO is a frequent adverse event after blood transfusion, and that it contributes greatly to the transfusion-related morbidity and mortality. Early reports mentioned in [9] indicated incidence values ranging from 40 to 1000 per million blood components transfused. Haemovigilance networks may provide useful information in this matter. Results from Québec [40,41] indicate a global incidence of 21 per million blood component in the years 2003–2004, while the French Hemovigilance observed an incidence of 10 per million blood component (Haemovigilance 2006 report) and the UK 3 per million blood component [42] (Table 2). The most consistent data seem to be those for fresh-frozen plasma (FFP), with three major observations: not a single case of TRALI has been observed using solvent/detergent plasma. When considering the French data alone, none of the 460,000 transfused solvent/detergent plasma provoked TRALI, while among the recipients of 650,000 quarantine FFP, 14 developed TRALI (P = 0·006); the incidences observed in the three countries in the absence of any preventive measure are very similar: 21 (France), 16 (Québec) and 16 (UK) per million FFP; and the implementation of a simple preventive measure in UK led in the years 2004–2006 to a dramatic reduction of TRALI incidence, to 3·2 per million FFP transfused. Results for red cell concentrate are more difficult to interpret, the incidences varying from 0·7 to 17 per million red cell concentrates. The very low incidence observed in UK may reflect on one hand the fact that Serious Hazards of Transfusion requires to declare serious adverse reactions, and it is known that many authentic TRALIs are not life-threatening, and on the other hand the fact that not all hospitals participate in the network, while the data provided in term of delivered blood components are national data. It is tempting to say that, provided that the notified cases are well-confirmed TRALI cases and that plasma content of red cell concentrate is similar in the different countries, the highest observed incidence, that is, 17 par million red cell concentrate in Québec is probably the closest to reality. Published results for platelets actually deserve clarification: France observed incidences of 8 and 50 per million platelet concentrates from whole blood and apheresis, respectively, while Québec declares 107 per million whole blood platelet concentrate, and Serious Hazards of Transfusion provides a global incidence of 14 per million platelet concentrate, without distinguishing between whole blood-derived and platelet The French results in the years to 2006 provide also information during four related to red cell concentrate transfusion and to platelet concentrate transfusion. Finally, one can a between the incidence of TRALI and the mean plasma volume from donors present in blood components as identified by the national control as in between the volume of plasma from donors present in the different blood components and the incidence of TRALI to the French Hemovigilance Apart from the investigations to the management of the and from the that be done in order to control the there is no specific to related to the blood component and the corresponding blood Once a TRALI is suspected, several investigations are mandatory in order to to detect an immunologic the transfused blood components and this information to the blood transfusion that have delivered patient's necessary for a and it to the laboratory in of the immunologic this not that all the investigations shall be but that the laboratory in of investigations shall be able to them if and all the donors and for the way to for class I and II and use stored plasma of the corresponding if and the donors to for investigations (in case many donors are potentially to identify if positive antibodies against class I and II and antibodies against neutrophils in the blood donors In case of any positive for the presence of the corresponding antigen on the patient's and a that in some such as neutrophil it may be complex to these tests due to technical In most donors who have been in a TRALI case due to the presence of or antibody recognizing a patient's antigen are from of blood for direct therapeutic In contrast, donors found for antibodies may as well as those with identified antibodies but recognized as not responsible for the TRALI analysis of risk factors in patients and analysis of fluid balance before transfusion are the means for TACO. can be as and blood The of the in UK to use donors only for providing direct therapeutic plasma has been a and the incidence of TRALI related to FFP in the with the 2004–2006 after implementation showed a dramatic reduction of TRALI 3·2 FFP (P = France in with a plasma may be from or without The is more complex for platelet on the of have specific of antibody in donors, but up to now, no developed a national for these Although mentioned, one should have in mind that red cell concentrate may cause TRALI, and in the French Hemovigilance red cell TRALI may be and overall for of the However, in the present the only preventive measure that to red cell concentrate is to from further donations those donors that were directly in a TRALI already mentioned that solvent/detergent plasma can be considered as of any known TRALI Therefore, its use can be an to plasma. Moreover, at least in the of French the incidence of TRALI is much with platelet concentrate than with of platelet concentrate, as shown in Table This observation, along with many others in favour of the use of platelet concentrate to the management of recipients, as it is already done in some implementation of platelet in platelet concentrate enables to the amount of plasma in each platelet It has been clearly as to the incidence of adverse reactions As many other in favour of such a we can that information its in TRALI incidence severity will in the few Finally, one can also the fact that the of of blood components before being transfused may favour the accumulation of active in blood components there is pressure to the of platelet concentrate up to days. However, we can more to the present days time and to develop the of platelet concentrate in order to manage blood of a measure with and potentially for the

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.