How do we treat? Clinical haemotherapy: platelet transfusion


  • H. Schrezenmeier,

    1. Institute of Clinical Transfusion Medicine and Immunogenetics Ulm, German Red Cross Blood Service Baden-Württemberg – Hessia and Institute of Transfusion Medicine, University of Ulm, Ulm, Germany
    Search for more papers by this author
  • B. Höchsmann,

    1. Institute of Clinical Transfusion Medicine and Immunogenetics Ulm, German Red Cross Blood Service Baden-Württemberg – Hessia and Institute of Transfusion Medicine, University of Ulm, Ulm, Germany
    Search for more papers by this author
  • M. Wiesneth

    1. Institute of Clinical Transfusion Medicine and Immunogenetics Ulm, German Red Cross Blood Service Baden-Württemberg – Hessia and Institute of Transfusion Medicine, University of Ulm, Ulm, Germany
    Search for more papers by this author

  • Conflicts of interest: To be confirmed.

  • 4A-MS

H. Schrezenmeier, Institute of Clinical Transfusion Medicine and Immunogenetics Ulm, German Red Cross Blood Service Baden-Württemberg – Hessia and Institute of Transfusion Medicine, University of Ulm, Ulm, Helmholtzstraße 10, 89081 Ulm, Germany


Platelet transfusions play an important role in prevention or treatment of bleeding in patients with thrombocytopenia or severely impaired platelet function. In clinical haemotherapy a number of decisions are necessary, including choice of the type of platelet concentrate, transfusion trigger for prophylactic platelet transfusions and the dose of platelet transfusion.

Usage of apheresis platelet concentrates and pooled whole blood-derived platelet concentrates varies greatly between countries and individual institutions. A clear advantage of apheresis concentrates can only be demonstrated in allosensitized patients with HLA- or HPA-antibodies who receive antigen-compatible apheresis platelet concentrates. We follow the recommendation to base the product choice mainly on availability and medical indication.

Recent data on therapeutic instead of prophylactic transfusions require a new position-fixing on when platelet transfusions should be given. Given the attempt to optimize patient safety and to avoid life-threatening bleeding complications the available data still suggest using a prophylactic approach in routine settings outside of clinical trials for treatment of thrombocytopenia in the context of chemotherapy and impaired platelet production.

Various studies addressed the effect of platelet dose in prophylactic transfusions in patients with hypoproliferative thrombocytopenia. We will review these studies and discuss implications for clinical practice.

Some patients fail to achieve the appropriate platelet count increase after transfusion. This is still a challenging situation for clinicians and transfusion services. Approaches to identify the cause of platelet refractoriness as well as therapeutic algorithms will be proposed.


Platelet transfusions are indicated for the prevention or treatment of bleeding in patients with thrombocytopenia or platelet dysfunction. Transfusion of platelet concentrates (PC) steadily increased over the recent years. More than 1·3 million platelet concentrates are transfused annually in Europe and more than 2·8 million in the United States [1,2]. Platelet concentrates available for transfusion are prepared either from whole-blood donations or by platelet apheresis procedures. Two methods of preparing platelet concentrates from whole blood are in use. The predominant method in Europe is the buffy coat (BC) method, and in the United States the platelet rich plasma (PRP) method [3]. The two methods mainly differ in the centrifugation steps and have been described in detail in a recent publication in the ISBT Science Series [4].

Platelet transfusion is an essential part of the overall therapeutic strategy in patients with hypoproliferative thrombocytopenia, e.g. in haematological malignancies, solid tumours, bone marrow failure syndromes, intensive chemotherapy and allogeneic or autologous stem cell transplantation. In this article, we will focus on the aspects important for the transfusing physician with regard to transfusion trigger, platelet dose, type of platelet component and ABO compatibility.

Transfusion trigger for prophylactic platelet transfusions in patients with hypoproliferative thrombocytopenia

For long time, platelets were administered prophylactically when the platelet count of a patient with hypoproliferative thrombocytopenia fell below 20 × 109/l. However, a number of retrospective studies and prospective randomized trials have demonstrated that the transfusion trigger for prophylactic transfusions can be safely decreased to 10 × 109/l in stable patients without additional risks of bleeding (Table 1) [5–13]. This transfusion trigger for prophylactic platelet transfusions has now mostly been adopted in clinical practice.

Table 1.   ’Transfusion trigger’ studies with comparison between a transfusion trigger of 10 × 109/l vs. 20 × 109/l platelet count for prophylactic platelet transfusion
Main DiagnosisDesignNo. of pts. 10*/20**Severe haemorrhage (%) 10/20Death because of haemorrhage 10/20Mean no. of Platelet transfusions 10/20Mean no. of RBC Transfusions 10/20Literature
  1. HSCT, haematopoietic stem cell transplantation; AML, acute myeloid leukaemia; AL, acute leukaemia;

  2. *10: Tranfusion trigger of 10 × 109/l platelet count

  3. **20: Tranfusion trigger of 20 × 109/l platelet count

  4. #Mean number of RBC transfusions/patient-day.

  5. Sum of APC and PPC.

  6. §Total no. of units in the first 100 days after transplant.

HSCTR78/8114/170/010·4 ± 17/10·2 ± 116·0/5·9Zumberg et al., 2002
AMLR135/12022/201/07·1 ± 4·6/9·0 ± 5·29·6 ± 5·2/9·1 ± 4·1Rebulla et al., 1997
AML 58/4718/170/018·4/30·27·3 (0–22)/8·3 (2–28)Wandt et al., 1998
ALR37/41 0/07 (5–11)/11 (6–15)11 (8–14)/10 (6–14)Heckman et al., 1997
HSCT 103/8713/143/454/73§ Gil-Fernandez et al., 1996
 64/7715/18 6·5/4·50·34/0·30#Lawrence et al. 2001
AML 21/2742/300/08·4 ± 5·3/8·5 ± 5·5 Navarro et al., 1998

However, it is important to emphasize that the transfusion trigger studies excluded patients with clinical conditions or concomitant therapies associated with increased risk of bleeding (Table 2). The morning platelet count is just one tessera in the overall picture of the patient. Monitoring of the patients for signs of bleeding and assessment of the overall clinical situation is very important [13–16]. As demonstrated in an exploratory analysis of a randomized trial, the presence of grade 1 bleeding on the previous day was associated with a 2·6 times increased risk of clinically significant bleeding [16]. A trigger of 20 × 109/l is now widely recommended in patients with increased risk of bleeding [17].

Table 2.   Risk factors for bleeding in patients with thrombocytopenia
• Infections
• Fever > 38°C
• Leucocytosis
• Tissue damage,
 e.g. necrotic lesions;
 Skin and/or mucosa lesions in graft-versus-host disease (GvHD);
 Mucositis Grade III/IV after chemotherapy or radiation
• Uraemia
• Rapid decline of platelet counts
• Clinical signs of bleeding (e.g. petechiae)
• History of bleeding
• Concomitant plasmatic coagulation disorders
• Platelet function defects
• Concomitant administration of anticoagulants
• Newly diagnosed AML t(15;17) – PML/RARA and variants
• Splenomegaly
• Treatment with antithymozyte globulin (ATG)
• Sinusoidal obstruction syndrome (SOS)/veno-occlusive disease (VOD)

Although haematology analysers provide reliable full blood counts, they are known to be inaccurate at enumerating platelets in severe thrombocytopenia. It has been demonstrated that haematology analysers tend to overestimate the platelet count, which would result in undertransfusion of platelets [18–20]. This re-emphasizes the need for external quality control to improve analyser calibration for samples with low platelet counts. Also, it hints to the importance of assessment of clinical signs of haemorrhage as predictor of bleeding [21].

Prophylactic vs. therapeutic transfusion

Recent studies have evaluated an exclusively therapeutic transfusion strategy [22–24]. In clinically stable patients, platelet transfusions were only used when relevant bleeding occurred (more than petechial). In a study of 140 autologous peripheral blood stem cell transplantations, this approach was considered as safe and feasible [22]. However, results in terms of bleeding complications and consumption of platelet concentrates were compared to historical controls [22]. Preliminary data are available from a prospective, randomized trial in patients after autologous stem cell transplantation and patients with intensive chemotherapy for acute myeloid leukaemia (AML) [23,24]. In the AML study, bleeding complications (more than petechial) were significantly increased in the therapeutic arm showing a 2·3 times higher risk compared to the prophylactic arm. Five minor and two major cerebral haemorrhages were registered in the therapeutic arm and none in the prophylactic group. Two patients in the therapeutic group died of cerebral haemorrhage. The authors conclude that an increased risk of fatal cerebral bleeding cannot be excluded by this study [24]. A larger study would be necessary to determine the final safety of an exclusively therapeutic transfusion strategy.

Taking into account the low residual risk of adverse events of platelet transfusions and the ongoing efforts to reduce these risks even more (e.g. bacterial screening; pathogen inactivation; and plasma-reduced concentrates), it must be discussed whether it is justified to put patients at risk of severe and potentially life-threatening bleeding when trying to further reduce the use of platelet transfusions. The real benefit for the patient is the pivotal question for a therapeutic transfusion strategy. Just the saving of platelet transfusions as a sole economic objective is not sufficient to justify an exclusively therapeutic transfusion strategy.

Platelet dose

The optimal dose of platelets in prophylactic transfusion is controversial. Most transfusion services provide PC units with about 3 × 1011 platelets as standard dose in current clinical practice for adults. In paediatric transfusion, a dose of 10 ml PC per kg body weight is recommended. In clinical practice, the dose of platelets transfused falls in a very wide range from about 0·17 × 109 up to about 29 × 109 per kg [25].

Several randomized studies compared different doses (‘high dose’, ‘standard dose’, ‘low dose’) [26–32]. However, it is not well defined which dose is ‘standard’, ‘high’ or ‘low’. The doses compared in clinical trials ranged from 1·5 × 1011 up to 11 × 1011. The absolute platelet increment, i.e. the difference between post- and pretransfusion platelet count, is increasing with increasing dose of transfused platelets [26–28,30–32]. Several studies suggested that higher doses of platelets result in prolonged transfusion intervals [26–28,30–32]. In routine practice, this might be in particular relevant to the management of outpatients. In one study, the overall number of platelets transfused was significantly higher in the high-dose group [32]. In contrast, another study could not demonstrate a significant difference as the higher platelet dose was compensated by a longer transfusion interval [30].

Also, the effect of platelet dose on bleeding in hypoproliferative thrombocytopenia is controversial: In a large cohort of 1272 patients, the Platelet Dose Trial (PLADO Trial) did not find a difference in the incidence of bleeding ≥ grade 2 between the dose groups [32]. In contrast, the SToP study was terminated early after enrolment of 118 patients because of grade 4 bleeding in 3 patients in the low-dose group [31].

Type of platelet product

There is an ongoing debate whether platelet concentrates prepared from either whole-blood donations using the buffy coat method (BC) or the PRP method or from plateletpheresis are superior [33]. In summary, there is some advantage of pooled platelets prepared by the BC method over the PRP platelet concentrates. However, the comparison between apheresis PCs and pooled BC-derived platelet concentrates suggests equivalence of the products in non-allosensitized recipients, regarding platelet increment, transfusion reactions, allosensitization and bacterial contamination [33]. A clear advantage of apheresis PCs can only be demonstrated in allosensitized patients with HLA and/or HPA antibodies who need human leukocyte antigen (HLA) antigen-compatible apheresis platelet concentrates. On this basis, it was recommended to base the product choice mainly on availability and medical indication [34]. The comparative studies were mainly based on surrogate parameters [35]. We are missing sufficient comparative data on prevention or treatment of bleeding or the interval between transfusions. Thus, from the point of clinical relevance the current evidence suggests that none of the products is superior. Further studies are needed, in particular prospective trials comparing the product types for clinically relevant endpoints. When deciding about the platelet supply strategy, every Blood Service has to take into consideration its own specific requirements and conditions, e.g. number of available whole-blood donations; number of refractory patients, number of immunocompromised patients requiring cytomegalovirus-negative products. The decision must balance donor safety and optimal use of the blood donors′ gift with an optimal management of the patients. Furthermore, the comparison between the product types needs regular reassessment because new developments like pathogen inactivation also have impact on the assessment of the various product types.

ABO compatibility and RhD compatibility

Transfusion of ABO-major incompatible platelet concentrates is associated with lower platelet increment and a significant higher rate of refractoriness when compared to ABO-identical platelet transfusion [36–39]. The detrimental effect of non-identical ABO platelet transfusion on posttransfusion increments seems to be cumulative: Increasing numbers of transfusions of ABO-incompatible plasma were associated with progressively poorer mean increments. The same was true for platelet ABO-incompatible transfusions. In contrast, increasing numbers of ABO-identical transfusions were not associated with poorer increments [40]. Repeated incompatible platelet transfusions can induce a significant increase in anti-A or anti-B isoagglutinin titres [37].

Several studies suggested that ABO matching of platelet transfusions even might have a beneficial effect on survival after cardiac surgery and in acute leukaemia and (in combination with leucoreduction of blood components) also in lymphoma [41]. In contrast, in a group of almost 1·700 patients undergoing cardiovascular surgery, no difference in adverse clinical outcome was observed between the two groups according to the ABO compatibility of the first platelet transfusion [42].

Transfusion of ABO minor incompatible platelet concentrates can cause acute haemolysis and has in particular been reported after transfusion of apheresis platelet concentrates with high isoagglutinin titres [43]. Therefore, whenever possible, ABO-identical platelet transfusions should be given. If ABO-identical platelets are not available, we first provide ABO major compatible platelet concentrates. Our practice is to avoid minor incompatible platelet transfusions in recipients with less than 25 kg body weight. An alternative is to give plasma-reduced platelet concentrates in additive solution or to measure isoagglutinin titres in apheresis platelets.

Platelets carry no Rh antigens, but platelet units may be contaminated by RhD-positive red-blood-cells. There is evidence that the minimum dose of red blood cells necessary for primary anti-D alloimmunization is as low as 30 μl. Studies after RhD-incompatible platelet transfusions reported anti-D frequencies from 0% to about 12% (see review in [43]). Overall, the frequency of patients who developed anti-D was about 6% [43]. This is much lower than observed after transfusion of RhD-incompatible red-blood-cell concentrates to RhD-negative subjects. The risk of anti-D alloimmunizataion might even be lower at present time because of the low content of red-blood-cells in platelet concentrates produced with up-to-date procedures. In addition, many patients receiving platelet transfusions experience immunosuppression caused by the underlying disease and/or the treatment. The current practice of restricting RhD-negative patients to receiving only platelets from RhD-negative donors is challenged by a recent report that could not find anti-D immunization in 104 RhD-negative recipients of apheresis platelet concentrates and considers prophylactic use of RhIG for RhD-positive apheresis platelet transfusions in RhD-negative patients to be unnecessary [44].

However, given that there is still a residual risk for anti-D alloimmunization and considering the potential severe consequences of anti-D formation, RhD-negative female recipients at child-bearing age or younger should not receive platelet concentrates from RhD-positive donors [17]. If such a transfusion cannot be avoided, intravenous RhIG for preventing anti-D alloimmunization should be considered [17,43].

Platelet transfusion refractoriness

There is no definitive agreement on the precise definition of platelet refractoriness. However, a 1-h corrected count increment (CCI) < 5–10 × 109/l or a percentage platelet recovery (PPR) < 20% suggests platelet refractoriness [45,46]. Also, approaches using posttransfusion counts more promptly, e.g. 10 -min posttransfusion, are used [47]. CCI and PPR calculations require a platelet count of the transfused platelet concentrate that is not routinely available. Therefore, it also has been suggested to simply look for increments of fewer than 10 × 109/l after transfusion of a platelet concentrate in an adult [48]. Diagnosis of refractoriness should only be made when an insufficient increment has been observed on at least two sequential transfusions of ABO-identical platelet concentrates 10 min to 1 h posttransfusion [45,48].

Platelet transfusion refractoriness can be caused by immune and non-immune factors. Non-immune causes include fever, infections, disseminated intravascular coagulopathy, splenomegaly, drugs, sinusoidal obstruction syndrome, graft-versus-host disease and bleeding [45].

Immune-mediated platelet refractoriness is due alloimmunization to HLA antigens or human platelet antigens (HPA). Alloimmunization to HLA antigens is more common than to HPA [49]. The introduction of leucodepletion reduced the incidence of HLA allosensitization and platelet refractoriness and non-immune causes of refractoriness became more common than immune-mediated causes [38,50]. More than half of the patients with HLA allosensitization do not present with platelet refractoriness.

The first step in the evaluation of suspected immune-mediated refractoriness is an HLA antibody screening test. However, results obtained with different tests can be discordant [51]. A panel reactive antibody (PRA) score > 20% suggests HLA alloimmunization as cause of refractoriness [52]. Patients should be supported with platelet concentrates with an HLA A/BU or BX match [53]. If the antibody specificity can be determined, the patient should at least be supported with antigen-negative platelets [54]. Another strategy is to transfuse crossmatch-negative platelets that might allow selection of effective platelet products even from donors with HLA-A or HLA-B-antigen mismatches [55–57].

Inadequate increment to platelet concentrates selected with these strategies should prompt screening for HPA antibodies. It is also important always to consider non-immune causes of the refractoriness.