Examining Bleeding Risk, Transfusion-related Complications, and Strategies to Reduce Transfusions in Lung Transplantation

Although transfusion can be lifesaving, it also is associated with increased hydrostatic forces in the pulmonary microcirculation, transfusion-related acute lung injury, and mortality. National hemovigilance systems show a relatively low but constant transfusion-related mortality rate, with a substantial proportion of those from pulmonary complications. In the United States, approximately 22 deaths were definite, probable, or possible of a total of 52 deaths reported to the Food and Drug Administration (Silver Spring, Maryland) of 6.6 million blood components transfused in the United States, while the risk of death is 1.57 per 100,000 components issued in the United Kingdom in 2022.

Transfusion of the donor or recipient in the preoperative phase of care has been linked to worse outcomes after lung transplant. Impaired graft survival is reported after massive transfusion of the donor  a potential result of major physiologic changes, including the coagulopathy of trauma, endothelial dysfunction, and metabolic changes. Further, transfusion is a known sensitizing event in the pretransplant recipient.  In one study of 12,283 patients, recipients transfused before the transplant episode were more likely to be sensitized to human leukocyte antigen and have difficulty identifying a match, have long waitlist times, and have high waitlist mortality.  Therefore, pretransplant transfusion should be avoided where possible, and importantly, if transfusion is needed, leukoreduced products should be used to decrease the risk of human leukocyte antigen alloimmunization.

Bleeding and transfusion in the recipient during lung transplantation are common. During the surgery, excessive bleeding during dissection and native pneumonectomy may be present due to vascularized adhesions or scar tissue from previous surgeries. There is expected blood loss at cannulation, de-airing of each new graft, and reperfusion, which can be anticipated. Attention to amounts of blood loss in the field and cell salvage collected throughout the surgery will reveal the rapidity of blood loss and need for urgent intervention with volume, while frequent blood gas monitoring and coagulation monitoring helps to assess the adequacy of resuscitation. A relative need for volume due to vasodilation may further be encountered during and after reperfusion. Bleeding may be surgical, but it may also be secondary to coagulopathy in the setting of a massive inflammatory response related to the recipient’s critical illness, degree of tissue injury from ischemia and reperfusion, and/or level of mechanical circulatory support and anticoagulation. Because the literature on lung transplantation is mostly limited to retrospective studies, it is difficult to know whether surgical trauma, inflammation, blood transfusions, or a combination of these variables are ultimately injurious to the allograft.

During lung transplantation surgery, transfusion, which is marker of preoperative anemia and blood loss, is correlated with early allograft injury and primary graft dysfunction, which affects 15 to 20% of recipients.  Although transfusion is repeatedly linked to severe primary graft dysfunction, this is difficult to clinically differentiate from other forms of transfusion-related lung injury or transfusion-associated circulatory overload.  Analyses of single-center and multicenter cohorts of lung transplant patients highlight the association between blood loss and transfusion with worse early allograft function, primary graft dysfunction, chronic rejection, and both early and late mortality.  Thus, correcting preoperative anemia and mitigating blood loss is critical to improve patient outcomes.

The most compelling evidence linking intraoperative transfusion to primary graft dysfunction was published by the multicenter Lung Transplant Outcomes Group (N = 1,255),  suggesting that large-volume transfusion of red cells (more than 1 l) and platelets during lung transplantation is independently associated with early graft dysfunction.  In one multicenter study of 729 registry patients that bled during lung transplantation, those receiving more than 4 units intraoperatively were 2.2 times more likely to have severe allograft injury within 72 h.  Seay et al.  examined transfusion strategies for lung transplantation and identified that patients receiving an increased plasma-to-erythrocyte unit ratio had an associated greater degree of primary graft dysfunction, although patients with increased ratios in this cohort also received a higher total transfusion volume.

Downstream effects of severe primary graft dysfunction include acute graft failure, increased length of stay, tracheostomy, postoperative mechanical circulatory support, critical illness, increased utilization of hospital resources, and ultimately long-term allograft dysfunction and worsened survival. In sum, developing strategies to reduce perioperative blood loss and the resulting transfusions for this patient population is an opportunity to potentially reduce primary graft dysfunction and improve outcomes after lung transplantation.

There may be downstream risks of sensitization and chronic graft rejection in the transplant recipient specific to the type of blood component transfused. Registry data are limited as they do not include information regarding the use of nonerythrocyte products, including platelets, cryoprecipitate, plasma, or factor concentrates. White blood cells are found in blood components such as packed red blood cells  and platelets and expose transfusion recipients to human leukocyte antigen. Although the majority of red cell and platelet products are leukoreduced to reduce the risk of forming antibodies to human leukocyte antigen as much as possible, soluble human leukocyte antigen class I and II and Fas ligand molecules are also present in most blood components. 

Antibodies to human leukocyte antigen may or may not also be antibodies to donor-specific antigens, formed in response to the donor organ. Persistent donor-specific antibodies are associated with chronic lung allograft dysfunction. Interestingly, a recent evaluation of 380 lung transplant recipients indicates a nonlinear relationship between the amount of blood components transfused and antibody development, and not all patients that develop antibodies to human leukocyte antigen within 12 months develop antibodies to donor-specific antigen. One study reported that platelet transfusions, and not plasma or erythrocyte units, had an associated risk of developing new antibodies to human leukocyte antigen (odds ratio [95% CI], 1.18 [1.02 to 1.36]; P = 0.025). Of unclear significance presented in the same study, compared with patients who never developed donor-specific antibodies or antibodies to human leukocyte antigen, cryoprecipitate transfusion was associated with donor-specific antibody development within the first year more than other components of blood (odds ratio [95% CI], 2.21 [1.32 to 3.69]; P = 0.002); this may reflect severe bleeding and difficult postoperative course. More studies are needed to understand the relationship between blood product administration, as well as additional compounding factors in the recipient and long-term outcomes.

Bleeding can be challenging to quantify during cardiothoracic surgery given the presence of cell salvage and the ability to recover heparinized blood into the cardiopulmonary bypass circuit. For lung transplantation, shed blood in the field during an off-pump procedure or on extracorporeal membrane oxygenation (ECMO) is salvaged, washed, and hemoconcentrated using intraoperative blood scavenging systems (e.g., cell saver), but the venous reservoir cannot be used for blood lost in the field unless on full cardiopulmonary bypass.

Two clinically useful ways of quantifying perioperative bleeding have been published by expert consensus within the last decade for cardiothoracic surgery. One recommended by the National Heart and Blood Institute (Bethesda, Maryland) quantifies the amount of blood products transfused over the first 5 perioperative days and suggests a further analysis of patients on ECMO using a five-point ordinal score of thrombosis and bleeding. The second, published in 2014 by Dyke et al., produced a fairly comprehensive definition of bleeding, which quantifies blood loss by the number of blood products transfused by type, use of factor concentrates, delayed chest closure, chest re-exploration, and thoracostomy tube output.

Thus far, published reports on blood loss in lung transplantation have not been quantified this way but instead mainly by the amount of erythrocyte units needed during the case. Due to the increased risks of primary graft dysfunction, definitions for what constitutes bleeding during lung transplant have been varied across publications. “Large volume” transfusion of red cells intraoperatively (more than 1 l) is one commonly selected description for clinically significant bleeding. In other single-center studies in lung transplantation, massive transfusion was defined as an intraoperative erythrocyte unit requirement more than 5 U, and another defined it as more than 10 U in 24 h.   A summary of published measures and definitions of bleeding applicable to lung transplantation can be seen in table 1.

While the debate is ongoing in cardiac surgery regarding whether a liberal or conservative strategy for transfusion is best it is difficult to extrapolate a clear threshold for transfusion from cardiac surgery to lung transplantation. For patients who are on ECMO, the Extracorporeal Life Support organization recently published guidelines for transfusion targets and anticoagulation, but these are limited in that they do not consider the management of active bleeding.

Lung transplant patients are often at greater risk of poor organ perfusion due to major, unpredictable, or constant bleeding during lung transplant surgery, with predisposing patient and procedural risk factors discussed in later sections. The empirical threshold for transfusion of red cells is between 7 (restrictive threshold) and 10 g/dl (liberal threshold); however, the decision to transfuse should not be made on counts alone. The International Society for Heart and Lung Transplantation (Chicago, Illinois) guidelines for anesthesia management suggest basing decisions to transfuse red cells based on several pieces of information including calculation of oxygen delivery, measures of oxygen delivery and consumption (e.g., mixed venous saturation), and markers of end-organ and tissue perfusion, taking into account the rate of blood loss. Specific transfusion targets for platelet, plasma, cryoprecipitate, and factor or fibrinogen concentrates based on platelet counts, factor levels, and fibrinogen levels, as well as anticoagulation targets, remain undetermined with much center-to-center variability. For lung transplant recipients at our centers, we modified a decision support tool used commonly for cardiac cases with targets and treatments agreed upon by anesthesiologists, perfusionists, intensivists, surgeons, hematologists, and the transfusion service to help anesthesia teams treat clinical bleeding based on the results of available point-of-care tests, product and factor concentrate availability, and the stage of surgery. For patients who remain on mechanical circulatory support, extreme caution should be taken in heparin reversal or overaggressive treatment of factor or platelet abnormalities due to the risk of thrombosis of indwelling cannulas or oxygenators, particularly if clinical bleeding is absent, the patient is prothrombotic, or there are areas of stasis in the circuit (e.g., switching cannulas or circuits).

The 2021 International Society for Heart and Lung Transplantation guidelines for selecting lung transplantation candidates recommend evaluating thrombocytopenia, leukopenia, anemia, and other hematologic abnormalities before transplant.  Further, they recommend extreme caution in considering the transplantation of recipients carrying excessively high risk for complications, including bleeding diathesis, thrombophilia, and bone marrow dysfunction, because of the challenges in managing bleeding during the procedure.

Preoperative assessment of potential coagulopathy begins with taking the patient’s bleeding history, followed by routine evaluation of hemoglobin levels, platelet count, and coagulation testing. More advanced testing and preoperative evaluation by hematology should be reserved for those with unexpected history or laboratory abnormalities or specific syndromes. Of note, some niche lung transplant patient populations who develop end-stage lung disease are predisposed to bleeding abnormalities that need to be managed individually depending on the deficiency. Examples include end-stage cystic fibrosis with acquired vitamin K deficiency, Hermansky–Pudlak syndrome, a combination of interstitial lung disease and platelet degranulation abnormalities, and pulmonary hypertension with acquired von Willebrand factor deficiency and/or acquired thrombocytopenia from inhaled pulmonary vasodilators.

Patients with end-stage pulmonary disease may also suffer from overall frailty and malnutrition, which should intervened upon as part of prehabilitation. Nutritional deficiencies in vitamin B₁₂, copper, vitamin C, and vitamin K can result in cytopenias and clotting factor deficiencies, which may be corrected in advance of transplant. Preoperative anemia, vitamin, and iron deficiency should be corrected preoperatively, if possible. Although of some benefit in cardiac surgery, consideration of intravenous iron therapy to improve hemoglobin levels is not well established or studied in lung transplantation. Erythropoietin has not been well studied in lung transplantation, with limitations including an unknown surgery date and cost of the drug relative to the time to clinical benefit. The risk and benefit profile is unclear, particularly for chronically hypoxic patients, who may already have high endogenous erythropoietin levels, or patients with pre-existing hypercoagulability. One small study examined whether thrombopoietin receptor agonists could be used in advance of lung transplantation to reduce platelet transfusion, but the results did not demonstrate a meaningful difference in the preoperative platelet count or the need to transfuse platelets during the transplant. 

Additionally, recipients may require systemic anticoagulation before transplantation, most commonly for atrial fibrillation, pulmonary embolism or deep-vein thrombosis, or antiplatelet agents for in-cardiac or vascular stents. A team-based programmatic discussion is required to identify which agents can be held at listing or easily managed and reversed once listed for transplantation. Although there is some center-to-center variability, vitamin K antagonists are frequently employed and easy to reverse with four-factor prothrombin complex concentrates during lung transplantation. 

As bridging patients to lung transplantation with ECMO becomes more common, recall that these patients have a potential bleeding diathesis in the form of vascular access for cannula or ECMO-related coagulopathy and paradoxically that some may be hypercoagulable (e.g., chronic thromboembolic disease). The ECMO circuit represents a prothrombotic foreign blood–surface interface. Consequently, these patients warrant early evaluation of the coagulation profile and anticoagulation. During initiation of ECMO, there is a consumptive coagulopathy and dilution of coagulation factors  as well as platelet activation and reduction of fibrinogen levels; this may in itself precipitate the need for transfusion.  The thrombogenicity of the circuit stabilizes over time, after platelet activation and absorption of fibrinogen to the circuit surface, with eventual replacement by other coagulation-neutral proteins.  Baseline viscoelastic testing in addition to traditional coagulation screen and platelet count before lung transplantation is helpful to identify the recipient’s coagulation profile, prepare targeted blood products, and discuss anticoagulation management.

Multiple patient factors are associated with increased risk of bleeding while undergoing cardiothoracic surgery; these included advanced age, chronic use of antiplatelet or anticoagulation, small body surface area, female sex, preoperative anemia, redo chest surgery, and elevated creatinine.  However, although there are some similarities, lung transplantation may be different enough from cardiac surgery that these risk factors may not all be directly extrapolated. The risk of bleeding and transfusion during lung transplantation are substantial, and several unique procedural risk factors are shown in table 2.  Massive transfusion during lung transplantation has an incidence of approximately 18 to 27% and is associated with a higher 90-day mortality. 

In a monocentric cohort of lung transplantation, transfusion was more frequent in cystic fibrosis and pulmonary fibrosis patients, potentially due to the need for mechanical support intraoperatively and/or a history of previous thoracic surgery in the pulmonary fibrosis cohort. Preoperative anemia, extremes of body mass index, and pleural adhesions secondary to chronic pleural inflammation in the cystic fibrosis group was also associated with transfusions. Notably, the increased allogeneic erythrocyte transfusion in the cystic fibrosis group may be related to the lack of cell salvage used in this patient population, which often has chronic infections and suppurative disease.

The risk of bleeding for lung transplantation is difficult to assess in national registries due to institutional and provider variability in mechanical circulatory support use, anticoagulation strategy, and transfusion practices. Specific components of mechanical circulatory support that influence coagulopathy development include anticoagulation, cannulation, hemodilution, platelet dysfunction, circuit or oxygenator inflammatory response, time on mechanical circulatory support, and potential for hypothermia.

The variability in the choice to employ mechanical circulatory support, cannulation sites, circuitry and oxygenators, mode, and anticoagulation strategy prohibits a broad recommendation of bleeding risk assessment based on one characteristic alone. Centers may opt for a bilateral and single lung transplantation approach, either on or off bypass, with or without veno-arterial or veno-venous ECMO or cardiopulmonary bypass. In one multicenter analysis of data from 11 national and international transplant centers, the mode of mechanical circulatory support influenced the outcome of primary graft dysfunction, with cardiopulmonary bypass having the highest risk of primary graft dysfunction and bleeding compared to ECMO and off-pump strategies.  This should be interpreted cautiously, since mechanical circulatory support can be used either electively or emergently when unexpected intraoperative complications lead to cardiopulmonary compromise. Because it is impossible to identify intent and rationale in retrospect, our definitive understanding of the relationships of bleeding, transfusion, primary graft dysfunction, and mortality continues to be challenged. Recently, there has been a surgical practice shift toward the use of elective veno-arterial ECMO during lung transplantation, with purported outcome benefits being controlled reperfusion and a lower targeted clotting time for anticoagulation. The bleeding risk associated with this specific strategy is unclear.

For patients requiring ECMO, targets and strategies for optimal anticoagulation remain undefined with higher levels of anticoagulation resulting in more bleeding and transfusion. For patients who are not bleeding on ECMO, lower therapeutic targets for anticoagulants and blood conservation are emphasized in a recent Canadian expert consensus document.  A single dose of 2,000 to 5,000 IU unfractionated heparin at cannulation, with monitoring of drug effect, is safely administered for lung transplantation on ECMO because typically the circuitry is heparin-coated. Recent guidelines from the American Association of Thoracic Surgery have suggested using a circuit without anticoagulation in patients who are at high risk of bleeding based on the Northwestern University (Evanston, Illinois) experience. Direct thrombin inhibitors including bivalirudin are increasingly used in ECMO as potential alternatives to unfractionated heparin. Regardless of anticoagulant choice, monitoring the drug effect by activated clotting time or partial thromboplastin time alone is not an adequate, comprehensive measure of the patient’s coagulation status, and additional measures of platelet count, fibrinogen, and thromboelastographic are helpful in measuring bleeding risk. 

Blood management strategies during the surgery include cell salvage; employing surgical maneuvers such as harmonic scalpel, argon beam, or electrocautery; minimally invasive surgical strategies; and low or minimal heparinization strategies to bleeding and transfusion risks, although currently the literature in this area is limited to retrospective observational trials and consensus documents. It is important to point out that for lung transplant patients without chronic infections, it is recommended to salvage blood from the field directly into the venous reservoir in a fully heparinized patient on cardiopulmonary bypass. However, during an off-pump or procedure supported by ECMO, there is no venous reservoir available, so blood lost to the field should be salvaged, centrifuged, and concentrated red cells administered back to the patient. This reduces the amount of allogeneic erythrocyte transfusion needed. However, when extreme amounts of cell saver have been processed, it may become necessary to replete clotting factors and platelets that are lost in the supernatant.

When patients bleed, a multidisciplinary, lab-based, targeted treatment approach has been shown to demonstrate success in reducing transfusions, morbidity, and costs overall. In cardiac surgery, implementing a bleeding management or transfusion algorithm reduced transfusions, transfusion-related complications, and cost. Although identifying universal transfusion thresholds based on counts is not realistic, the use of point-of-care testing as a measure of abnormalities may improve defining indications for transfusion.

Use of factor and fibrinogen concentrates as alternatives to allogeneic blood products to manage coagulopathy varies highly between US and European centers, largely due to cost and availability. In the United States, some centers have started to report use of factor concentrates for lung transplantation, specifically prothrombin complex concentrates in place of plasma for correction of clotting factors without increasing thromboembolic risk or factor VIIa use for rescue of bleeding in lung transplant patients, which has thrombotic risk.  Fibrinogen concentrate is available to treat bleeding due to hypofibrinogenemia in lieu of cryoprecipitate. Major considerations for their use should be availability, dose relative to deficiency, cost, and timing of treatment relative to mechanical circulatory support.

In many European countries, aprotinin is approved, available, and recommended for use in lung transplantation.  Fibrinolytics reduce bleeding and have been shown to have transfusion-sparing effects without substantial thromboembolic risk in both cardiac and noncardiac surgery. However, other fibrinolytics such as tranexamic acid and aminocaproic acid have not been studied well in lung transplantation and have not been studied in the context of performing the procedure on or off ECMO with little to no heparin.  Within our authorship group representing one US and one European center, one uses antifibrinolytics routinely in all lung transplant cases, and the other uses antifibrinolytics prophylactically for cardiopulmonary bypass cases or for the treatment of bleeding during an off-pump or ECMO procedure. An informal survey of US and international centers demonstrates a wide practice variability between providers and lung transplant centers.

Although point-of-care coagulation testing or target-based therapies are used extensively in cardiac surgical patients, few studies are available in lung transplantation. Smith et al.  compared a point-of-care coagulation testing–specific algorithm based on rotational thromboelastometry and platelet aggregometry assays to a standard of care in a retrospective cohort of 93 patients undergoing lung transplantation on cardiopulmonary bypass. In the point-of-care coagulation testing–based strategy group, transfusions were reduced but without significant differences in outcomes; a cost analysis reported a reduction in transfusion costs. In a single-center randomized clinical trial, Durila et al.  compared rotational thromboelastometry–guided transfusion to a standard of care in lung transplantation that included two fluid resuscitation strategies: albumin 5% in point-of-care coagulation testing group versus crystalloids in non–point-of-care coagulation testing. The study showed a reduction in perioperative bleeding and transfusions in the point-of-care coagulation testing group, potentially due to the provider using data from both point-of-care coagulation testing and use of albumin over crystalloids. Other viscoelastic tests are currently available and widely reported in cardiac surgical patients. An ongoing multicentric randomized controlled trial (NCT05798286) with point-of-care coagulation testing thromboelastography may offer better guidance of coagulation strategy and treatment of coagulopathy in lung transplantation.

Reducing bleeding and transfusion is a significant challenge for clinicians as increasingly complex patients present for lung transplantation. Understanding recipient risk and implications of transfusions to the recipient are important to optimize strategies for reducing bleeding, transfusions, and improving outcomes. Specific topics of further study include optimization of the donor and recipient, surgical approaches to lung transplant and cardiopulmonary support, use of more consistent and comprehensive measures of bleeding in this population for improved comparisons of registry data, and strategies to reduce transfusion and achieve hemostasis.