Guidelines for the use of platelet transfusions

The demand for platelets in England was stable at around 220 000 adult therapeutic doses (ATD) per year until 2007/8 at which point demand has increased year-on-year to 275 000 ATD in 2014/15, an increase of 25%. Similar rises in demand have been seen in Australia and the United States. A recent review which considered causes for this dramatic rise identified that an ageing population and an increase in the incidence of haematological malignancies (with increased treatment intensity, duration and survival) accounted for most of this change (Estcourt, 2014). In 2012 the population in the UK aged over 70 years was 7·5 million. By 2046 this number is expected to reach 15 million (Office of National Statistics (ONS) 2013). In addition, since 1990, the number of haematopoietic stem cell transplants performed in Europe has risen, from 4200 to over 30 000 annually (Passweg et al, 2012). Although a national audit of platelet use in haematology identified that 28% of transfusions were outside of guidelines (Estcourt et al, 2012a), these findings demonstrate less inappropriate use than a previous audit (Qureshi et al, 2007). An increase in the proportion of inappropriate use is therefore unlikely to have contributed significantly to recent changes in demand (Estcourt, 2014). Currently up to 67% of all platelets are used in the management of patients with haematological malignancies (Cameron et al, 2007; Greeno et al, 2007; Pendry & Davies, 2011; Jones et al, 2013; Charlton et al, 2014). Much of the remainder are used in cardiac surgery (7–10%) and in intensive care (5–9%). In contrast to platelet demand, the donor base is steadily dropping, with a 35% reduction in active donors from 1·893 million in 2000 to 1·231 million in 2015 (NHS Blood & Transplant, unpublished data). As the majority of platelets in the UK are collected from approximately 14 000 registered platelet donors (apheresis platelets), and whole blood donors give blood on average 1·7 times a year this could have a significant impact on the future supply (European Blood Alliance 2015, European Committee (Partial Agreement) on Blood Transfusion CD-P-TS 2016). This guideline aims to provide practical advice on platelet transfusions to help clinicians to decide when support is expected to be beneficial and to reduce inappropriate use. If the reason for thrombocytopenia is unclear, further investigation is required as this is likely to influence management. This document will cover practice in adults relevant to the UK and replace the 2003 British Committee for Standards in Haematology (BCSH) platelet use guideline (British Committee for Standards in Haematology Blood Transfusion Task Force 2003). A one page summary document is available in Appendix 1. The indications for platelet transfusion in children and neonates and more general specifications, such as cytomegalovirus (CMV) status and irradiation, are not included, and can be found elsewhere (New et al, 2016; Advisory Committee on the Safety of Blood, Tissues and Organs (SaBTO) 2012, Treleaven et al, 2011). The classification of platelet transfusion into either 'therapeutic', to treat bleeding, or 'prophylactic', to prevent bleeding, was based on the modified World Health Organization (WHO) bleeding score (Table 1) (Stanworth et al, 2013a). Recommendations for prophylactic transfusion relate to patients with bleeding scores of 0 or 1 and therapeutic transfusion to patients with bleeding scores of 2 or higher. For each indication, the recommendations include a threshold or target platelet count and a suggested dose, when relevant. The Grading of Recommendations, Assessment, Development and Evaluation (GRADE) nomenclature [Audit tool (Appendix S1)] was used to evaluate levels of evidence and to assess the strength of recommendations. The GRADE criteria are specified on the BCSH website (http://www.bcshguidelines.com/BCSH_PROCESS/42_EVIDENCE_LEVELS_AND_GRADES_OF_RECOMMENDATION.html) and the GRADE working group website (http://www.gradeworkinggroup.org). A search of published literature was undertaken using the evidence from several systematic reviews that are either currently being undertaken by members of the writing group (Estcourt et al, 2014a,b,c), or that have been recently published (Hedges et al, 2007; van Veen et al, 2010; Lieberman et al, 2013; Pavenski et al, 2013; Wardrop et al, 2013; Kumar et al, 2014; Vassallo et al, 2014; Crighton et al, 2015; Nahirniak et al, 2015). This was supplemented by searching PubMed and the United Kingdom Blood Transfusion Services (UKBTS)/Systematic Review Initiative (SRI) Transfusion Evidence Library (www.transfusionevidencelibrary.com) up to November 2014 using specific search terms relevant to each section of the guidelines. The guideline group was selected to be representative of UK-based medical (anaesthetics, benign and malignant haematology, haemostasis, transfusion) and laboratory experts with practical experience in platelet transfusion. Given the breadth of application, the draft guideline was provided to sounding board members of the Haemato-oncology, General Haematology, Haemostasis and Thrombosis, and Transfusion Task Forces of the BCSH for comment and subsequent revision. Recommendations for Prophylactic Transfusion of Platelets to Patients with Thrombocytopenia Because Of Reversible Bone Marrow Failure Where Recovery Is Anticipated: Recommendations for Prophylactic Transfusion of Platelets to Patients with Thrombocytopenia Because Of Chronic Bone Marrow Failure, Where Recovery Is Not Anticipated: Recommendations for Prophylactic Transfusion of Platelets to Other Patient Groups: Recommendations for Prophylactic Platelet Transfusion Prior To Procedures or Surgery: Recommendations for Therapeutic Platelet Transfusions: Recommendations for Platelet Function Disorders (Congenital): Recommendations for Platelet Function Disorders (Acquired): Recommendations for Immune Thrombocytopenia (ITP): Contraindications to Platelet Transfusions: Risks from Platelet Transfusions: Recommendations for Platelet Refractoriness: Recommendations for Other Alternatives or Additions to Platelet Transfusion: The evidence for these recommendations is based on studies in patients with haematological malignancy causing thrombocytopenia due to the disease or its treatment. Other patient populations are considered separately. A systematic review identified six randomised controlled trials (RCTs) that compared a prophylactic versus therapeutic platelet transfusion strategy (Crighton et al, 2015). Four of the included studies were conducted at least 30 years ago and used out-dated methods of platelet component production and patient supportive care. Two of the included studies were recent large RCTs (Wandt et al, 2012; Stanworth et al, 2013a), both of which showed that prophylactic platelet transfusions reduced the risk of bleeding when all patients with haematological malignancies receiving treatment (e.g. chemotherapy or transplantation) were considered (Crighton et al, 2015), but this effect was not seen in a pre-specified sub-group ‒ patients receiving autologous haematopoietic stem cell transplants (HSCT) (Table 2) (Stanworth et al, 2014). This finding indicates that prophylactic transfusion should continue to be the standard of care in patients receiving intensive chemotherapy or allogeneic transplantation but may not be appropriate in low risk groups with short periods of thrombocytopenia. A systematic review identified three RCTs that compared different platelet transfusion thresholds (Estcourt et al, 2012b). Two compared a threshold of 20 × 109/l vs. 10 × 109/l, whereas the third compared a threshold of 30 × 109/l vs. 10 × 109/l. A fourth RCT excluded from the systematic review compared a threshold of 20 × 109/l vs. 10 × 109/l (Zumberg et al, 2002). A meta-analysis of all four studies (658 patients) showed that the 10 × 109/l threshold was not associated with increased bleeding in comparison with a higher threshold and also showed a significant reduction in the number of platelet transfusions given (Estcourt et al, 2011). However, this meta-analysis may not be sufficiently powered to detect an increased bleeding risk in this lower threshold arm of less than 50% (Estcourt et al, 2011). The use of other transfusion thresholds, such as platelet mass, absolute immature platelet numbers and immature platelet fraction, have been considered as alternatives to a platelet count threshold but there have been no randomised studies in adult patients. (Eldor et al, 1982; Briggs et al, 2006; Gerday et al, 2009; Zisk et al, 2013). A systematic review identified six RCTs that compared different platelet transfusion doses (Estcourt et al, 2015). Four of these studies assessed clinically significant bleeding as an outcome measure (usually defined as WHO grade 2 or above). There was no evidence of a difference in the risk of bleeding between low dose (1·1 × 1011/m2) and standard dose (2·2 × 1011/m2) and between standard dose and high dose platelet transfusions (4·4 × 1011/m2). Low dose transfusions decreased the total amount of platelets patients received, but at the expense of a higher number of transfusions episodes. Increasing the dose from a standard to a high dose did not increase the transfusion interval (median 5 days for both regimens). The mean UK adult platelet dose (one unit of platelets) is around 3 × 1011 platelets, equivalent to between the low and standard doses defined above, although there is evidence of considerable variation (Pietersz et al, 2012). Numerous clinical factors have been reported to be associated with an increased risk of bleeding (Table 3). However, the majority of these postulated risk factors are based on low-level evidence, such as expert opinion or retrospective analysis of patient databases. Inflammation has been shown to be associated with an increased risk of bleeding in mice (Goerge et al, 2008). Although studies have differed in their opinion of whether fever increases the risk of bleeding in humans (Table 3), currently, the platelet transfusion threshold is commonly raised to 20 × 109/l when patients have an infection or fever (Estcourt et al, 2012a). Further studies are required to clearly identify which factors should prompt an increase in the transfusion threshold, and what this threshold should be. r = 0·174 (major haemorrhage) P < 0·001 r = 0·054 (minor haemorrhage) P < 0·001 r = 0·064 (major haemorrhage) P < 0·001 r = 0·114 (minor haemorrhage) P < 0·001 r = −0·101 (major haemorrhage) P < 0·001 r = −0·144 (minor haemorrhage) P < 0·001 OR 1·14 (95% C.I. 0·89–1·46) (not transfused) OR 0·98 (95% C.I. 0·69–1·39) (transfused) Other co-morbidities Splenomegaly Other co-morbidities Recent history of severe haemorrhage (≤5 days) There is little evidence to inform practice. A retrospective study considered platelet transfusion in outpatients with stable chronic severe aplastic anaemia (AA) (Sagmeister et al, 1999). Prophylactic platelets were given if the count was 5 × 109/l or less. In total, 55 239 patient days were reviewed of which there were 18 706 days when the platelet count was 10 × 109/l or less. All deaths from haemorrhage were associated with alloimmunisation or withdrawal from treatment. Three non-fatal major bleeding episodes occurred. The authors concluded that this restrictive policy, with a median transfusion interval of 7 days, was feasible, safe and economical. International guidelines that consider patients with chronic thrombocytopenia recommend either a 'no prophylaxis' strategy (Schiffer et al, 2001; Liumbruno et al, 2009; Kaufman et al, 2015) or prophylaxis below a count of 5 × 109/l (The Board of the German Medical Association on the Recommendation of the Scientific Advisory Board 2009). A major concern in using a threshold of 5 × 109/l is the reported inaccuracy of current automated counters when the platelet count is very low (Segal et al, 2005; De la Salle et al, 2012). A policy of prophylaxis has an impact on resources and on patient quality of life. Recent BCSH guidelines for the diagnosis and management of adult AA and for the diagnosis and management of adult myelodysplastic syndromes (Killick et al, 2014, 2015) advise a no prophylaxis strategy for patients who are not receiving active treatment, with the latter including patients taking low dose oral chemotherapy or azacitidine (Killick et al, 2014). Platelet function defects, immune-mediated thrombocytopenia and thrombotic thrombocytopenic purpura are considered in later sections. There is little evidence to guide practice in other patient populations. One patient group who are significant users of prophylactic platelet transfusions are those in critical care. A large observational study of critically ill patients showed that 9% (169/1923) of all critically ill patients received platelet transfusions and 55% (296/534 units) of these were given on days when no significant bleeding occurred (Stanworth et al, 2013b). The optimal platelet transfusion management of these patients (Lieberman et al, 2013) may differ depending upon the underlying clinical diagnosis (Assir et al, 2013). As the evidence base in non-haematological patients is sparse we have extrapolated the evidence from studies in haematology patients to this population as a basis for our recommendation until further evidence is available. According to the confidential registry of complications after bone marrow aspirates and trephines the risk of significant bleeding is very low (less than 1 in 1000), and the majority of patients with bleeding did not have significant thrombocytopenia (Table 4). Maintaining pressure on the biopsy site until bleeding has stopped is advised. Seventeen observational studies have reported bleeding outcomes in thrombocytopenic patients after insertion of central venous catheters (CVCs) (Table 5). Only one case of severe bleeding (Hb drop >15 g/l) was reported throughout all of these studies (Weigand et al, 2009). Three studies reported on risk factors, in addition to thrombocytopenia, associated with bleeding. In two of these studies ultrasound guidance was not used and, on multivariate analysis, the risk of bleeding was significantly increased by the number of attempts, site of insertion (jugular versus subclavian) and failed guide-wire insertion (Barrera et al, 1996; Fisher & Mutimer, 1999). In the third study, where ultrasound guidance was used, no such correlation was identified (Zeidler et al, 2011). Systematic reviews of complications of CVC placement (Randolph et al, 1996; Hind et al, 2003) and a more recent small study (Tomoyose et al, 2013) found that ultrasound guidance significantly reduced failure and complication rates. Observational Retrospective 115 (NR) 0 (0) Observational Retrospective 1825 (2010) INR > 1·3 aPTT > 37 s 3 (0) Observational Retrospective 40 (259) PT <40% aPTT ≥ 77 s 0 (0) Observational Prospective 115 (115) 20 (0) Observational Prospective 1660 (1978) Observational Retrospective 0 (0) 76 (104) 7 (0) Observational Prospective 283 (658) 19 (0) Observational Retrospective 2514 (3170) 0 (0) Observational Retrospective 80 (80) 0 (0) Observational Prospective 105 (112) 0 (0) 133 (133) NR (0) 72 (108) 4 (0) Observational Prospective 196 (NR) NR (1) Observational Retrospective 193 (604) 5 (0) Observational Retrospective 55 (57) 10 (0) Observational Retrospective 1 (0) Zeidler et al (2011) looked at the risk of bleeding according to platelet count thresholds with multivariate analysis. All CVCs were un-tunnelled and inserted by experienced individuals and the analysis was controlled for sex, type of leukaemia, insertion site and use of prophylactic platelet transfusions. The risk of bleeding only increased when the platelet count was less than 20 × 109/l (Odds ratio 2·88, 95% confidence interval 1·23–6·75, P = 0·015) (Zeidler et al, 2011). In a large study by Haas et al (2010), tunnelled CVCs were installed and all bleeding episodes were effectively controlled by simple pressure at the site of insertion. The platelet count threshold for insertion was 25 × 109/l (Haas et al, 2010). One additional prospective study assessed insertion of peripherally inserted central catheters (PICCs) without prophylactic platelet transfusions (Potet et al, 2013). Among the 50 patients who had a line inserted with a platelet count less than 20 × 109/l, only one bleeding episode occurred (minor oozing). One prospective non-randomised study assessed the risk of bleeding after traction removal of tunnelled cuffed CVCs in patients with abnormal platelet counts or an increased International Normalised Ratio (INR) (Stecker et al, 2007). Of the 179 patients enrolled in the study, 14 had a time to haemostasis of over 5 min and only one of these patients had a platelet count <100 × 109/l. A wide-ranging review of the literature has been performed to assess the risk of spinal haematoma following lumbar puncture and spinal and epidural anaesthesia. The evidence was based on case series, case reports and expert opinion. There was insufficient information to consider epidural and spinal anaesthesia separately (van Veen et al, 2010). The authors recommend that providing the platelet count is stable and no additional coagulopathy or platelet function defect is present a platelet count of ≥80 × 109/l should be used for placing/removing an epidural catheter or performing spinal anaesthesia and a count of ≥40 × 109/l for lumbar puncture (van Veen et al, 2010). As the technique for spinal anaesthesia is comparable to that of a lumbar puncture, a count of ≥40 × 109/l for both of these procedures and a separate threshold of 80 × 109/l for epidural anaesthesia would be more logical. We are aware of no new studies that have contributed to the literature since this review. A total of 2740 percutaneous liver biopsies were conducted in the Hepatitis C Antiviral Long-term Treatment against Cirrhosis (HALT-C) trial (Seeff et al, 2010); only 16 patients (0·6%) had a serious adverse event due to bleeding. Percutaneous liver biopsies are considered safe when the platelet count is at least 50–60 × 109/l (British Society of Gastroenterologists (BSG) 2004, Rockey et al, 2009). Below this level, transjugular liver biopsy (TJLB) should be considered. This procedure has been shown to be safe in patients with low platelet counts and with modern techniques can produce comparable histological samples to those from a percutaneous route (Wallace et al, 2003; Kalambokis et al, 2007; Mammen et al, 2008). Patients with uraemia have a platelet dysfunction that is thought to be associated with von Willebrand factor (Hedges et al, 2007). Uncontrolled hypertension, high serum creatinine, anaemia, older age and female sex have been shown to be risk factors for bleeding following renal biopsy and to prolong the bleeding time (Manno et al, 2004; Whittier, 2004; Torres Munoz et al, 2011; Zhu et al, 2014). Reversal of these problems by treatment of hypertension (Zhu et al, 2014), dialysis (Hedges et al, 2007; Mannucci, 2012), the use of desmopressin (Mannucci et al, 1983; Hedges et al, 2007; Manno et al, 2011) or conjugated oestrogens (Mannucci, 2012) and the correction of anaemia (Hedges et al, 2007) have all been reported to reduce the risk of bleeding in non-RCTs. Although treatment of anaemia with recombinant human erythropoietin can take many weeks a more rapid effect on haemostasis has been noted. This may be through improved platelet and et al, et al, and an increase in the number of platelets 7 days et al, renal biopsy has been used in patients in percutaneous renal biopsy has failed or been and has a and et al, Platelet transfusion is likely to be at very as the dysfunction the platelets will be The transfusion may also be in patients who to renal of the risk of alloimmunisation et al, 2013). One recent small RCT patients procedures and has shown a low of bleeding complications without blood in patients to liver transplantation et al, 2012). Patients had platelet counts × 109/l, an INR and were randomised to the or of on used to A third of patients had a platelet count × 109/l. Only one patient in the arm had bleeding, which was controlled with Further is required a recommendation can be to use There a of evidence to guide the prophylactic use of platelet transfusions major from around the a threshold of 50 × 109/l major surgery (British Committee for Standards in Haematology Blood Transfusion Task Force et al, 2006; Liumbruno et al, 2011; Vassallo et al, and a threshold of × 109/l to or surgery the of the of the critical (British Committee for Standards in Haematology Blood Transfusion Task Force et al, 2006; Liumbruno et al, 2011; Vassallo et al, 2013). surgery is an procedure and therefore platelet transfusions are not of the platelet count following platelet transfusion is but may be by the There is little evidence for the of platelet transfusions or the optimal dose when a patient with thrombocytopenia is bleeding WHO grade 2 or (Estcourt et al, 2013). This may the in trials in these clinical and also the that platelet dysfunction may with major bleeding that is not by the platelet count et al, 2012). One recent large national audit reported the of bleeding after a therapeutic platelet transfusion in of with clinically significant bleeding grade 2 or (Estcourt et al, 2012a). recommendations are based on guidelines from around the (British Committee for Standards in Haematology Blood Transfusion Task Force et al, 2006; et al, 2010; Liumbruno et al, National Blood 2011; et al, 2013; Vassallo et al, 2013) and recently BCSH guidelines for major haemorrhage et al, 2015). is a severe bleeding in which many patients not on the platelet This increases the risk of alloimmunisation to platelet and to platelet which may prevent the treatment of bleeding et al, 2006; et al, 2008). is as a treatment in patients with and platelet However, most UK experts also line for the treatment or of bleeding in and that platelet transfusion should be considered for bleeding or high surgery et al, In less severe platelet function including and desmopressin may be for If a patient has recently an platelets to or the of of the will the defect as the platelets et al, 2013) (Table The effect of platelet transfusion to bleeding outside of this critical is A recent RCT of adults taking with haemorrhage found no evidence to support the use of platelet transfusions et al, 2016). This the findings of two systematic reviews the treatment of adults on with or haemorrhage which found no evidence of a all included studies were of low or very low quality & 2012; et al, 2012). In and a case that platelet dysfunction by is to with platelet transfusion than treatment with or et al, 2007; et al, 2012; et al, 2014; et al, 2015). A study in 14 these in as two of platelets was shown to low platelet but there was no in platelet et al, 2011). with 2 with Platelet transfusion to the of is although increases the risk of bleeding not increase bleeding for most procedures et al, 2013). In addition to of platelet transfusion for other than many patients who are these are at high risk of and a platelet transfusion may increase this risk et al, 2013). In the RCT of haemorrhage in patients on et al, as as no evidence of the of or at 3 were higher in those who received platelet transfusion compared to those who received standard care. In a study of 14 patients two of platelets to surgery to the of and one patient 4 days after surgery and had been and after et al, 2012). In has been used in three RCTs in patients taking (with or without patients) and significantly reduced blood transfusion et al, 2012; et al, in adverse were reported between the the authors the small numbers and the two studies was for 1 of the effect of uraemia is in the section on Platelet transfusion to procedures and surgery is an immune-mediated by thrombocytopenia count <100 × in the of underlying et al, 2010). and patients have little or no bleeding, whereas can experience Platelet transfusions are not as prophylaxis et al, 2010). The Association advise that for and thrombocytopenia, if the patient and platelet count are stable and the can be when the count is × 109/l when performed by a and experienced & 2010). Platelets have been used, in with other to treat major bleeding et al, 2011). There are no of case observational studies and A review of these studies (Table that or platelet transfusions have been at bleeding, if the platelet count has not been A platelet count rise to be more if platelet transfusions and are et al, & and one study this is more clinically & 2008). after × time to time to bleeding 1 unit 30 min × 1011 per (1) (1) 2 (1) bleeding improved bleeding decreased or stopped to over unit unit on the diagnosis and management of have been published et al, 2012). has been that a platelet transfusion may increase the risk of & 2011; et al, 2012). However, the evidence for this is and based on two case from the patients in et al, et al, Two more recent case patients in have reported no with & et al, 2010). This is a associated with severe thrombocytopenia following blood transfusion and by against can be serious and The incidence has reduced since the of are the