• blood transfusion;
  • diagnosis;
  • prevention;
  • side effects


  1. Top of page
  2. Abstract
  3. Introduction
  4. Acute non-infectious side effects
  5. Delayed non-infectious side-effects
  6. Disclosures
  7. References

As infectious complications from blood transfusion have decreased because of sophisticated blood screening, non-infectious side-effects have emerged as the most common complications of transfusion in industrialized countries. Despite making every endeavour, quite a number of these side-effects are very difficult to control. Some of these side-effects present as acute transfusion reactions, the most important of which are transfusion-related acute lung injury, circulatory overload, sepsis, and allergic and anaphylactic reactions. Acute adverse events require immediate action, but are often difficult to evaluate. On the other hand, transfusion recipients may experience delayed non-infectious side effects, including, non-ABO haemolytic transfusion reactions, graft-versus-host disease, and post-transfusion purpura. In addition, a not well-defined risk of increased mortality following blood transfusion has been recognized. Some aspects regarding the clinical presentation, the physiology behind, and possible preventive measures are summarized here.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Acute non-infectious side effects
  5. Delayed non-infectious side-effects
  6. Disclosures
  7. References

The use of blood products is not without hazards, and side-effects are seen in a relevant number of transfusions. Side-effects are defined as problems that occur in addition to the intended therapeutic effect, or that occur when treatment goes beyond the intended effect.

On the one hand, side-effects can be based on the presence of wanted or unwanted (contaminating) substances in the blood product. As an example, (wanted) red-blood-cells may lead to alloimmunization, and unwanted (contaminating) viruses may lead to infection. On the other hand, side-effects can occur as a consequence of the incorrect use of blood components, in which ‘incorrect’ can be either obvious and avoidable or masked and not (or not always) avoidable. As an example, red-blood-cells may lead to acute haemolysis in ABO-incompatible transfusion (avoidable), but the transfusion of two units of compatible red-blood-cells may lead to volume overload and lung oedema (not always avoidable).

Over the past decade, concern regarding side-effects associated with the transfusion of blood products has shifted from infectious disease transmission to non-infectious side-effects. In industrialized countries, a patient is 1000-fold more likely to experience a non-infectious side-effect than an infectious complication of transfusion. Many of these non-infectious side-effects are currently not or not completely avoidable. These side-effects are in the focus of this overview.

Acute non-infectious side effects

  1. Top of page
  2. Abstract
  3. Introduction
  4. Acute non-infectious side effects
  5. Delayed non-infectious side-effects
  6. Disclosures
  7. References

Some side-effects present as acute transfusion reactions within minutes or hours after transfusion including, transfusion-related acute lung injury (TRALI), transfusion-associated overload (TACO), transfusion-associated sepsis (TAS) and allergic/anaphylactic reactions. These acute transfusion reactions can be difficult to evaluate. The fact that pulmonary manifestations may dominate the clinical features further impedes the differential diagnosis.

Transfusion-related acute lung injury

TRALI is a serious, often life-threatening pulmonary transfusion reaction characterized by non-cardiogenic lung oedema, hypoxemia and respiratory distress in temporal association with blood transfusion. Since Barnard’s initial description in 1951 [1], non-cardiogenic lung oedema related to transfusion has been widely reported. A broader recognition, however, was gained only after publication of the results of the British ‘Serious Hazards of Transfusion’ (SHOT) initiative [2] which has consistently demonstrated since 1996 that TRALI is one of the most common causes of transfusion-associated major morbidity and death.

The initial symptoms of TRALI are caused by the onset of pulmonary oedema. Respiratory distress, hypotension, fever and cyanosis appearing within minutes to just a few hours from the initiation of the relevant blood transfusion are typical symptoms of TRALI (Table 1). In ventilated patients, a sudden drop of arterial oxygen tension may occur, and copious frothy oedema can ooze from the endotracheal tube. Consensus criteria for the diagnosis of TRALI were established in 2004 [3]. The incidence per unit transfused has been calculated as 1:75 000–1:88 000 for plasma and platelets and as 1:557 000 for red-blood-cells [4]. The real incidence of TRALI has not yet been established.

Table 1.   Differential diagnosis of acute non-infectious side-effects
ParameterTransfusion-related acute lung injuryTransfusion-associated overloadTransfusion-associated sepsisaAllergic/anaphylacticb
  1. ainitial assessment (septic shock is a dynamic process).

  2. bnot anaphylactic shock.

Clinical findings
 Body temperatureFever may be presentUnchangedFeverUnchanged
 Blood pressureHypotension, may be severeHypertension (post-transfusion systolic blood pressure elevation > 30 mmHg)HypotensionHypotension may be present
 RespirationAcute dyspneaAcute dyspneaDyspnea may be presentDyspnea may be present
 Neck veinsUnchangedDistension may be presentUnchangedUnchanged
 SkinUnchangedUnchangedUnchangedFlushing, hives may be present
 GastrointestinalNoneNoneNausea, vomitingNausea, vomiting, abdominal cramps, diarrhoea
 AuscultationCrackles, paucity of findingsCrackles, rales (S3 gallop may be present)+/−Stridor may be present
 Fluid balance+/−Positive+/−+/−
 Response to diureticsMinimal, sometimes deteriorationSignificantNoneNone
Additional findings
 Chest radiographNew bilateral infiltratesNew bilateral (central) infiltrates, enlarged cardiac silhouette, Kerley’s B linesUnchangedUnchanged
 EchocardiographyNormal or decreased ejection fractionDecreased ejection fractionEjection fraction may be decreasedNormal
 Pulmonary artery occlusion pressure< 18 mmHg> 18 mmHg<> 18 mmHg< 18 mmHg
 Central venous pressureUnchangedIncreasedUnchanged or decreasedUnchanged
 Oedema fluidExudateTransudateNoneNone
Laboratory findings
 WBCTransient leukopenia may be presentUnchangedLeukocytosis or leukopenia may be presentUnchanged
 B-type natriuretic peptide< 100–200 pg/ml> 500–1200 pg/ml+/−< 100–200 pg/ml

Prospective trials of different treatments have not been reported so far. In mild reactions, oxygen support can be sufficient whereas artificial ventilation is required in severe TRALI. Arterial hypotension can often be managed with intravenous fluid; diuretic treatment may worsen the clinical condition and is not indicated. There is no convincing data to support or refute the use of corticosteroids or, as recently suggested, aspirin [5]. Typical episodes improve within 48 h; the mortality rates are in the range of 1–10% [4].

In the majority of cases, TRALI induction can be attributed to the passive transfusion of leucocyte antibodies present in blood components [6]. These antibodies are directed against human neutrophil antigens [6] and human leucocyte antigens (HLA) class I and class II [7,8]; they are mainly detected in blood components donated by (multi-)parous women. By different mechanisms, these antibodies lead to stiffening of the neutrophil membrane, entrapping of the rigid neutrophils in the pulmonary capillaries, and, finally, to the disturbance of the pulmonary blood-air barrier, with the induction of lung oedema. Antibodies alone are not always capable of inducing TRALI but may require the presence of further triggers, as illustrated in the ‘threshold model’ of TRALI (Fig. 1). Deferral of female donors or of female donors with a history of pregnancy is a promising way to reduce the incidence of TRALI significantly [4]. TRALI has also been reported after the transfusion of neutrophil-priming substances accumulated in stored cellular blood components (including, bioactive lipids and CD40L). Non-immune TRALI is characterized by a more benign clinical course, and preventive measures have not yet been established. Such measures could include washing of cellular blood components or reduced shelf-life of products, but this awaits clinical studies.


Figure 1.  In the threshold model the occurrence of transfusion-related acute lung injury (TRALI) and its severity depends on the degree of neutrophil activation by the transfused TRALI mediators and the patient’s individual susceptibility to TRALI. Accordingly, where the external stimulus to neutrophil activation is sufficient, lung damage can occur in otherwise healthy individuals with no other likely cause of lung injury. However, these cases are the minority, because most patients receiving transfusions have significant comorbidities, which may also affect the lung tissues. Early reports noted that most patients with TRALI had recently undergone surgery. The proposed threshold model of TRALI summarizes these factors as the strength of transfusion-related mediators (antibodies, lipids, etc.) on the one hand and as the individual predisposition of the transfusion recipient on the other hand, covering both constitutive (genetic) and comorbidity-related factors (infection, disseminated intravascular coagulation, aspiration, shock etc.). Depending on the magnitude of the neutrophil response, lung injury can be mild or severe with corresponding clinical effects. Adopted from [6].

Download figure to PowerPoint

Transfusion-associated circulatory overload (TACO)

TACO has long been known as a complication of massive plasma transfusion as required in the treatment of patients with thrombotic thrombocytopenic purpura. With the broader recognition of TRALI as a dangerous complication of blood transfusion, TACO was also more often reported and a number of studies have improved our knowledge during the last years.

Most patients with TACO develop respiratory distress within 1–2h of transfusion. They frequently complain about chest tightness and headache, and a dry cough is common. Tachypnea, tachycardia, cyanosis and an elevated blood pressure are clinical signs of TACO (Table 1). SHOT has developed a diagnostic score for TACO [2]. Chest radiography, echocardiography, pulmonary artery occlusion pressure and B-type natriuretic peptide levels may be helpful in diagnosing TACO and distinguishing it from TRALI [8]. The incidence of TACO is not easy to evaluate because worsening of cardiac function or coincident disruption of the underlying fluid balance are difficult to differentiate from excessive transfusion. It can be expected to be in the range of 1000–6000 cases per 100 000 transfused patients, depending on age and comorbidity [8].

The first steps in treatment are upright positioning and supplemental oxygen. If symptoms persist, diuretics will be required; some patients may require intubation and positive pressure ventilation. TACO is not a benign condition: the reported fatality rate is between 3.6% and 20%.

The basic physiology behind TACO is that if the amount of blood in the pulmonary vessels is increased, regular transvascular fluid filtration is elevated and the fluid removal by the lymphatics is overcharged, which leads to lung oedema. There is evidence that many patients with TACO had excess fluid before and suffer from congestive heart or renal failure, all of which reduces volume tolerance [9]. Volume overload is best prevented by close attention to the state of the patient’s circulation. Slow transfusion and diuretics may minimize the risk of overloading.

Transfusion-associated sepsis (TAS)

Bacterial contamination of blood components results from the introduction of bacteria at the time of phlebotomy or, less commonly, from asymptomatic donor bacteremia or during blood banking procedures. Despite the fact that bacteria can be grown from a substantial number of blood components (especially, platelet concentrates), clinical sepsis as a consequence of blood transfusion is relatively uncommon.

Typical clinical features are fever, chills, hypotension and nausea (Table 1). Progress to septic shock may occur. There is no specific clinical marker for TAS, and the diagnosis may depend on the results from bacterial cultures taken from the patient and the component. It can be assumed that up to 1:1000 red-blood-cell concentrates (RBC) are contaminated with bacteria and as many as 1:100 platelet concentrates (PC). Platelet pools have a 5.6-fold higher risk of bacterial contamination than single-donor PCs [10]. The incidence of transfusion reactions caused by bacteria is estimated at 1:100 000 to 1:1 000 000 for RBCs and 1:900 to 1:100 000 for PCs. Serious complications are more often after the transfusion of contaminated RBCs [11].

Septic shock requires intensive care treatment modified according to the clinical presentation of the patient. Mortality rates in the United States caused by TAS are estimated at 1:20 000 to 1:85 000 [12].

As outlined earlier, bacteria that enter the storage bag because of asymptomatic donor bacteremia, contamination during collection (especially from skin) or blood banking procedures will grow while the blood product is stored. The clinical relevance of this contamination depends on the amount of bacteria present at the time of transfusion, their type and pathogenicity, including, their ability to produce endotoxins and may range from no effect to lethal septic shock. Exclusion of donors with diarrhoea or dental procedures, skin disinfection in two phases, predonation sampling, testing of tube connections for tightness and good manufacturing practice may prevent bacterial contamination of blood components. Some countries have established methods to detect bacterial contamination [12], produce single-donor PCs preferably, or have reduced the shelf-life of blood products. Pathogen inactivation techniques are a promising preventive alternative [12].

Allergic and anaphylactic transfusion reactions

Allergic and anaphylactic reactions are often rapid and potentially dangerous reactions to an immunologically foreign substance in a sensitized person.

Allergic reactions are characterized by skin flushing, urticaria, pruritus, and, less common, facial, laryngeal and pharyngeal oedematous swelling. In major allergic (anaphylactic) reactions, hypotension, bronchospasm and stridor as well as gastrointestinal symptoms can be present. In most patients, the diagnosis is easy to make, even if pulmonary symptoms are leading (Table 1). Anaphylactic transfusion reactions occur in approximately 1:50 000 transfusions, but PCs are five times (and plasma is two times) more often involved in allergic reactions than RBCs [13]. Mild allergic reactions can be seen in 1% to 5% of all platelet transfusions.

Antihistamines and corticosteroids are effective in treating allergic reactions; severe allergic reactions may require airway and circulatory support.

IgA deficiency is one cause of anaphylaxis that is discussed intensively in the literature, but less than 20% of patients experiencing a severe allergic reaction are immunized against IgA [14]. Anaphylactic reactions have also been associated with anti-Chido/Rodgers, anticomplement, antihaptoglobin and HLA antibodies; it appears, however, that the vast majority of reactions are not related to these antibody entities. It has also been speculated that specific substances released from platelets (such as microparticles) could be responsible for a relevant number of allergic reactions [13]. Patients with a history of severe allergic reactions and proven sensitization against IgA may require washed PCs and RBCs. Because the relevant allergen is unknown in the majority of patients with repeated allergic reactions, pretransfusion treatment with H1 receptor antagonists and/or corticosteroids should be considered.

Delayed non-infectious side-effects

  1. Top of page
  2. Abstract
  3. Introduction
  4. Acute non-infectious side effects
  5. Delayed non-infectious side-effects
  6. Disclosures
  7. References

Non-ABO haemolytic transfusion reactions (HTR), transfusion-associated graft-versus-host disease (TA-GVHD) and post-transfusion purpura (PTP) do usually not manifest before days or weeks after transfusion and present with more or less distinct clinical picture that will allow for a definite diagnosis. Furthermore, blood transfusion appears to carry a not well-defined risk of increased mortality [transfusion-associated mortality (TAM)] that cannot be attributed to any of these side-effects. TAM is been suspected to be associated with red-blood-cell storage lesions and with transfusion-related immunomodulation potentially promoting cancer recurrence and perioperative infections.

Non-ABO haemolytic transfusion reactions (HTR)

Despite dozens of mismatched major antigens between donor and recipient in every RBC transfusion, less than 10% of chronically transfused patients develop RBC antibodies. The recipient’s immune status and the dose and frequency of transfusions are factors that may influence the rate of alloimmunization. A non-ABO acute haemolytic transfusion reaction can occur if a RBC alloantibody is (accidentally) disregarded prior to transfusion, e.g. in an emergency situation. This, compared to ABO-HTRs, is a rare situation (17% of all deaths related to acute HTR reported to the US FDA) [15]. It is well known that alloantibodies may escape detection weeks or months after they were formed because of a rapid decrease in their titre. Exposing the patient again to the antigen that had induced antibody formation will be followed by an anamnestic immune response, and the antibody will become detectable again. This is called a delayed serological transfusion reaction. In delayed HTR (DHTR), the recalled antibody has the capability of inducing haemolysis.

Anaemia and jaundice are frequently seen in patients with delayed HTR; clinical signs commence within 14 days after transfusion. Acute haemolysis with haemoglobinuria and renal failure is rare but has been reported. The diagnosis can be made by a positive direct Coomb’s test; the alloantibody can be eluted from the RBCs.

There is no specific treatment for DHTR. Patients may require additional RBC transfusions. Severe reactions are very rare, and the overall prognosis is excellent.

As outlined earlier, in delayed HTR, donor RBCs are destroyed 7–10 days after transfusion, following an anamnestic immune response to a donor RBC antigen to which the recipient has been alloimmunized before. In most patients, these antibodies are of anti-c or anti-Jka specificity. Haemolysis is of the extravascular type in nearly all patients, i.e. the cells are removed by macrophages in the spleen and the liver. Intravascular haemolysis with a severe clinical course is very rare in non-ABO HTR. There is no specific prevention for DHTR. RBC antibodies that have been detected in the patient’s serum should be considered for all future transfusions.

Transfusion-associated graft-versus-host disease (TA-GVHD)

Graft-versus-host disease has been a known complication of allogeneic bone marrow transplantation since the 1960s. It was not until the 1970s that TA-GVHD was recognized as a distinct entity. Since then, it has been reported in a variety of patient groups, and various risk factors have been identified (Table 2) [16].

Table 2.   Transfusion-associated GVHD: risk factors and irradiation recommendations for cellular blood productsa
Irradiation recommended Risks definedNo irradiation recommended
Risks under discussionNo risks defined
  1. aadopted from [16]

  2. bIn allogeneic BMT or stem cell transplantation, irradiation is recommended for products given up to 14 days before and up to 6 months after transplantation, or until immune system reconstitution; in autologous BMT or stem cell transplantation, irradiation is recommended for products given up to 14 days before and up to 3 months after transplantation.

  3. cIrradiation is recommended for blood products given to patients in all stages of their disease.

Transfusion from blood relatives Human leucocyte antigens-matched transfusions Granulocyte transfusions Intrauterine transfusions Exchange transfusions Congenital immunodeficiency BMT and stem cell transplantationb Hodgkin’s diseasec NHLc Therapy with purine analogue drugsHematologic malignancies Solid tumours Solid organ transplantationTerm and preterm infants AIDS

Clinical features include, fever, rash, diarrhoea and liver dysfunction occurring at about day 10 after transfusion. Bone marrow failure with pancytopenia will develop subsequently. The disease is more progressive than GVHD after bone marrow/stem cell transplantation, often resulting in the death of the patient within 3 weeks after transfusion. In newborns, onset and course of TA-GVHD are somewhat delayed. The definite diagnosis is made by skin histology and the documentation of the presence of donor-derived cells, or DNA, in the blood or affected tissues of the patient. The overall incidence of TA-GVHD is difficult to assess but appears to be in the range of 1:1 000 000.

There is no effective treatment of TA-GVHD. Several immunosuppressive regimens have been reported, and some patients were successfully treated with stem cell transplantation. However, the overall mortality of TA-GVHD is still > 90% [16].

The pathogenetic cell behind TA-GVHD is the viable lymphocyte contained in cellular blood components. Transfused lymphocytes often remain detectable in the circulation of the transfused recipient before the immune system of the recipient removes them. As a prerequisite for TA-GVHD, transfused donor lymphocytes have to engraft and proliferate. This may occur if the donor is homozygous for an HLA class I haplotype of the recipient or, if the recognition and elimination of donor lymphocytes is inadequate because of a compromised immune system. It is currently unknown in which way and to what extent the immune system has to be affected to put the patient ‘at risk’ for developing TA-GVHD. Patients with different diseases have been recognized as being at risk (Table 2). When identified, these patients should be provided with gamma-irradiated cellular blood components only. In contrast, irradiation of plasma does not appear to be mandatory. Leucocyte reduction of cellular components alone cannot be used as an exclusive method to prevent TA-GVHD [2]. In contrast, photochemical treatment may have the potential to prevent TA-GVHD.

Post-transfusion purpura (PTP)

Post-transfusion purpura is a rare clinical syndrome characterized by the occurrence of severe thrombocytopenia, usually occurring 5–10 days after blood transfusion. Typical cases of PTP are observed in women with a history of pregnancy, who are likely to be immunized against human platelet alloantigens (HPA), mostly HPA-1a. PTP can also occur in male patients where primary immunization has occurred by blood transfusion.

Approximately 1 week after blood transfusion, an acute, severe thrombocytopenia is observed. Platelet counts may fall below 10G/l. Haemorrhage (cutaneous bleeding, melena, haematuria) is common and may be severe; even intracranial bleeding has been reported. The diagnosis is obvious in patients with acute, isolated, severe thrombocytopenia in temporal association with blood transfusion, but may be complicated in patients with more complex underlying disease. The diagnosis of PTP is confirmed by the demonstration of alloantibodies in the patient’s serum against one of the HPA antigens. The incidence of PTP with leucodepleted blood products is estimated as less than 1:700 000 [2].

Treatment of choice is the infusion of high-dose immunoglobulins (2 g/kg body weight) over two consecutive days, divided into 2 doses per day. A response rate of 85% has been reported. Platelet transfusions are not effective [17].

The majority of patients with PTP were immunized against HPAs (mostly HPA-1a) in a previous pregnancy or, less common, a previous blood transfusion. A recent blood transfusion triggers this anamnestic immune response. Thrombocytopenia is caused by antibody-mediated destruction of both donor platelets and the patient’s own platelets. The mechanism by which autologous platelets are destroyed is not completely understood. It appears that during the recall response, for unknown reasons, self-reactive antibodies are formed. Autologous platelets could also be removed as ‘innocent bystanders’ if they become coated with membrane fragments or immune complexes formed as a consequence of the immune reaction against the donor platelets. General leucodepletion is effective in preventing PTP; in the United Kingdom, the number of PTP cases has dropped from 10/year to 2/year after introduction of leucodepletion [2]. In patients with a history of PTP, further PC transfusions should be matched for the respective HPA.

Transfusion-associated mortality (TAM)

Blood transfusion appears to carry a not well-defined risk of increased mortality that cannot be attributed to any of the hitherto mentioned side-effects.

In several studies, transfusion of ‘older’ RBCs (compared with ‘fresh’ RBCs) has been associated with increased mortality, prolonged hospitalization, intensive care treatment, mechanical ventilation; an increased risk of postoperative pneumonia, infection at any site, and multiorgan failure. Only a few randomized clinical trials were performed, which, in summary, do currently not support a clear association between the age of RBCs and any of the mentioned side-effects. In a recent meta-analysis (as-treated), transfusion of old RBCs was even associated with a significant reduction of in-hospital mortality [18]. In addition, a large number of observational studies were performed in trauma, intensive care unit (ICU), cardiac surgery and colorectal surgery patients. Analysing these data reveals no clear trend towards increased mortality, organ failure, infection, length of stay in hospital or ICU care [18]. One important reason for this is that the significance of most studies suffers from not adjusting the data for the number of units transfused. Patients in the ‘old’ RBCs arm have often received more RBCs in average compared with recipients of ‘fresh’ RBCs. The number of RBCs transfused reflects more severe illness, comorbidity, and poorer baseline prognosis [18]. Taken together, the suspicion that ‘old’ RBCs are associated with common adverse morbidity/mortality outcomes is currently not supported.

Besides length of storage, white-blood-cell contamination has been associated with increased TAM as a consequence of transfusion-related immunomodulation. The idea that leucocytes in transfused RBCs may lead to immunomodulation stems originally from observational studies revealing better survival of kidney grafts in previously transfused patients compared to non-transfused patients. Based on these findings, cancer growth and impaired immunity against infections were suspected consequences of blood transfusion in many publications. However, cancer growth was not found to be influenced by transfusion of leucodepleted and non-depleted RBCs in two randomized studies [19]. In addition, an association between leucocyte-containing transfusions and the incidence of postoperative infections was not supported by a recent meta-analysis [20]. Even the graft-tolerizing effect of pretransplant allogeneic leucocytes in blood products could not be reproduced in more timely studies [19].

In contrast to these findings, an increased postoperative mortality has consistently been established for patients with open heart surgery when comparing leuco-reduced vs. non-reduced transfusions in five randomized trials. This increased risk could be associated with factors prevalent in cardiac surgery patients such as systemic inflammatory response syndrome where contaminating leucocytes may set a second insult in addition to the surgical procedure [19]. The detailed pathogenesis remains to be identified. There is no specific treatment or intervention available, but transfusion of leucodepleted RBCs will be helpful in preventing immunomodulatory effects in recipients.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Acute non-infectious side effects
  5. Delayed non-infectious side-effects
  6. Disclosures
  7. References