|
American Red Cross Biomedical Services
Medical and Scientific Updates Transfusion of Platelets: Current Issues
Gary Moroff, Ph.D. Prepared and Distributed by Holland Laboratories and the Medical Office Introduction The use of platelets has increased substantially over the last 15-20 years, in large measure because of the enhanced supportive care required by patients with malignancies and the use of stem cell transplants. In a recent review, Murphy suggested that utilization may be plateauing (1). This may reflect a lower utilization per admission, perhaps due to more rigorous evaluation of when and how much to transfuse. However, with an aging population and an increase in aggressive medical practices, the total number of patients receiving platelets should be increasing. The potential for associated adverse reactions, as reviewed in a recent editorial by Kruskall, continues to be of concern (2). Use of pre-storage leukoreduced platelets, whether prepared by apheresis or filtration, is reducing the risk of occurrence of specific adverse reactions. Although the use of platelet transfusions has resulted in substantial knowledge about when and how to support thrombocytopenic patients, there are still many questions about the practices that are routinely used. In addition, there is an interest in having a better understanding of the true costs associated with platelet utilization. This report will, therefore, review information and opinions about the following issues:
Types of patients receiving platelets Although it has been assumed that hematology/oncology patients use 70-80% of transfused platelets, there is limited documentation that this is correct. In fact, Petz reported at the 1997 AABB Meeting that at the UCLA Medical Center 48.2% of platelets are used by the hematology/oncology service, 15.4% by cardiothoracic surgery, 22.5% by liver transplantation and 13.9% by other services (3). In addition, this report stated that "nationwide statistics" suggest that about 40% are used by hematology/oncology, 21% by cardiovascular surgery, 3% by liver transplantation and 36% by other services such as general surgery and trauma. There is limited documentation in the literature pertaining to platelet utilization in association with organ transplants, other than with liver replacement procedures. Triulzi and Griffith recently published an analysis of data from primary lung transplantation at the University of Pittsburgh Medical Center in 1994 and 1995 (4). Thirty six (36)% and 4% of double and single-lung transplants required platelets. The average amount utilized was the same for single and double lung recipients who underwent cardiopulmonary by-pass procedures. Prophylactic and therapeutic use: targeted platelet levels at the time of transfusion Although there has been some controversy about the value of prophylactic therapy, most platelet transfusions for patients being treated for a malignancy or following a stem cell transplantation continue to be of this type. The primary question is at what level or threshold should such transfusions be administered. Historically, 20 x 109/L has been the accepted level. Over the last 5-10 years, it has been shown that many patients, especially those who are stable or without complicating clinical factors such as fever, do not have to be given transfusions until their platelet levels are reduced to 10 x 109/L, and in some instances to 5 x 109/L. The rationale for using a level as low as 5 x 109/L is based on Slichter’s observation that gastrointestinal bleeding in clinically stable patients was not substantially increased until the platelet count dropped to this level (5). Guidelines have been developed for treating patients with and without risk factors, utilizing different approaches. Gmur published a set of guidelines, termed "stringent", which specify levels for transfusion based on studies that his group has conducted (6). The guidelines include the following provisos: use of a 5 x 109/L threshold for patients without fever or bleeding; use of a 10 x 109/L threshold for patients with recent bleeding or fever; and use of a threshold of at least 20 x 109/L for patients with coagulation disorders or anatomical lesions or those on heparin. In 1996, AuBuchon reported that at his institution, the adoption of the Gmur criteria had resulted in a 52% reduction in platelet usage (7). Schiffer has presented the approach utilized at the University of Maryland Cancer Center which includes flexibility to individualize platelet transfusion decisions according to each patient’s condition (8). In that system, patients are grouped into categories. One group with long-term, nonreversible thrombocytopenia do not routinely receive platelets, regardless of platelet count, unless complications such as bleeding occur. Patients in a second group are given transfusions to maintain their platelet counts substantially above 20 x 109/L when there is an increased risk of bleeding because of the clinical condition or treatment. The third category, representing the largest group of patients, are considered for platelets on the basis of individual factors such as the presence of fever, infection and the likelihood of endogenous recovery of bone marrow function. Within this category, patients do not receive platelets if they are stable and bone marrow recovery is imminent. Many of these patients are likely to have a platelet count below 20 x 109/L. On the other hand, platelets are transfused at levels above 20 x 109/L if the patients have just received chemotherapy, have a falling platelet count, and are expected to have a lower platelet count the next day absent transfusion. The use of various thresholds continues to be a subject for clinical studies. A recent abstract by Hersh and Brecher indicated that at their medical center there was an increased bleeding risk with the use of a 5 x 109/ L threshold in bone marrow transplant patients with uncomplicated thrombocytopenia (9). One other recently published report summarized the data from a relatively large randomized study comparing the response to transfusions with thresholds of 20 x 109/L and 10 x 109/L (10). Newly diagnosed acute leukemia patients, treated at 21 medical centers in Italy, were randomly assigned to receive platelets when their platelet count fell below 20 x 109/L (120 patients) or less than 10 x 109/ L [or 10-20 x 109/L when the body temperature exceeded 38°C, there was active bleeding or invasive procedures were needed] (135 patients). The risk of major bleeding during induction chemotherapy was similar for the two groups of patients. Major bleeding, defined as any bleeding more than petechiae, mucosal, or retinal bleeding, was observed with 20 and 10 x 109/ L thresholds in21.5 and 20 percent of the patients. Such bleeding was observed on 2.0% and 3.1% of the days of hospitalization with the higher and lower thresholds, respectively. One fatal cerebral hemorrhage occurred with the lower threshold. Twenty one (21) percent fewer platelets were transfused using the 10 x 109/L threshold. The therapeutic use of platelets has not received as much published scrutiny as has prophylactic transfusion. It is generally assumed that increasing platelet counts to 40 - 50 x 1010 /L stops major bleeding episodes. There are limited data suggesting that when a platelet transfusion was given only when bleeding occurred, the number of serious bleeding episodes is not increased compared to that observed using prophylactic transfusions. The results of two studies indicated similiar mortality and red cell requirements with therapeutic and prophylactic usage (11). The Transfusion Practices Committee of the AABB has analyzed platelet transfusion practices based on data from a national survey conducted in 1992. The data showed that the majority of institutions (>70%) transfused platelets primarily in a prophylactic manner and that 60% of hospitals used a platelet count threshold of 20 x 109/L (12). In addition, the most frequently used target platelet count for performance of minor invasive procedures was 50 x 109/L. Changes in practices have occurred at many hospitals since 1992. However, the Committee’s 1995 published report on the 1992 study provides background on the variability of platelet transfusion practices. Platelet dose The dose of platelets to be transfused has traditionally been designated in terms of a specific number of whole blood-derived platelet concentrates or a single apheresis unit, without much regard for the number of platelets therein. It has been assumed that whole blood-derived platelet concentrates contain at least 5.5 x 1010 platelets. In a rather arbitrary way, 6 or 8 platelet concentrates are being pooled as a transfusion dose. Apheresis units have been assumed to contain at least 3.0 x 1011 platelets. These levels have been shown to provide a satisfactory clinical dose. With the increased harvesting of platelets from units of whole blood to levels of approximately 8-10 x 1010 per concentrate, some hospitals have reduced the number of platelet concentrates that are pooled to as few as 4 or 5 units (13 ). This "platelet unit approach" assumes that a platelet transfusion consists of an average number of platelets. A refinement to this approach stipulates that the number of platelets to be transfused should be based on the body weight of a patient. Although clearly the case when pediatric transfusions are given, the concept of utilizing an adult patients’s body weight to determine the number of platelet concentrates to be given is apparently considered by only a limited number of medical centers. Over the last few years, the advantages and disadvantages of the transfusion of high and low levels of platelets have been considered in terms of platelet survival kinetics and the period needed between transfusions. Herman (14) and Noral (15) have collected data indicating benefits from transfusing high levels of platelets. Herman concluded that higher platelet doses result in higher corrected count increments (CCI) and longer intervals between transfusions. The response to two dose levels of apheresis platelets with median values of 3.2 and 4.9 x 1011 platelets were compared in the prospective, randomized, multiple-crossover studies leading to this conclusion. The platelets were transfused to bone marrow or stem cell transplant recipients. On the other hand, at the 1997 AABB Meeting, Hersh and Brecher (9, 16) presented the concept that platelet survival is better when small, frequent doses of platelets are utilized as opposed to large doses. This concept was derived from a mathematical model that considered platelet survival as a function of changing platelet count. Patient data on the number of platelet concentrates needed to maintain a specific platelet count level in uncomplicated thrombocytopenia were used to validate the model. The lowering of the platelet dose from 6 to 3 platelet concentrates was associated with a reduction in the number of platelet concentrates needed. It was suggested that consideration should be given to the more frequent use of 3-unit platelet pools and one-half apheresis units. With the interest in reducing costs associated with the transfusion of platelets, issues pertaining to "acceptable" and "appropriate" dose levels will, no doubt, continue to be the subject of extensive evaluation and discourse. Assessment of platelet transfusion responses The response to prophylactic platelet transfusions is evaluated by determining the CCI. A one-hour CCI of 5-7.5 x 109/L, determined between 10 and 60 minutes post-transfusion, is considered to indicate a satisfactory response. A recent report emphasized that the reliability of the CCI measurement is dependent upon the accurate and timely assessment of the pre- and post-transfusion platelet counts (17). The report presented the results of studies that evaluated the reduction in platelet counts in 30 stem cell recipients with relatively uncomplicated thrombocytopenia. A major conclusion of the studies was that if the pretreatment count is obtained far in advance of the transfusion, and if the platelet counts falls markedly in the interim, then the CCI will be spuriously low. Results from the TRAP Study The results from the TRAP study, which evaluated the influence of leukocyte reduction and ultraviolet radiation as means to prevent alloimmunization to platelet transfusions, were published in December, 1997 (18). The study was a multi-institutional, randomized, blinded trial involving patients receiving induction chemotherapy for acute myeloid leukemia. Four study groups were utilized, each receiving a different type of platelet component. The control group received pooled whole blood-derived platelet concentrates that were not treated to reduce the leukocyte level. The three experimental groups received (a) filtered, pooled whole blood-derived platelet concentrates, or (b) ultraviolet B-irradiated pooled platelet concentrates, or (c) filtered, apheresis platelets. In total, 530 patients provided evaluable data, with approximately 130 patients in each study group. It was concluded that all three experimental platelet components are "equally effective in preventing alloantibody-mediated refractoriness to platelets during chemotherapy for acute myeloid leukemia". Quantitatively, 13% of patients in the control group developed lymphocytotoxic antibodies and become refractory to platelet transfusions, while 3-5% of patients enrolled in the experimental groups were observed to develop such characteristics. Platelet matching Alloimmunization against HLA antigens is the predominant immunological reason for refractoriness. However, in some instances, alloantibodies to platelet-specific antibodies may also be a cause of refractoriness. Such antibodies may explain refractoriness observed with HLA-matched platelet and some patients have been shown to have both types of antibodies. The successful treatment of refractoriness in six patients, four with acute myeloid leukemia and two with a myelodysplastic syndrome, who exhibited both types of antibodies, simultaneously, was summarized in a recent report (19). The six patients were documented to have alloantibodies to HLA and to either platelet specific antigen HPA-1b or HPA-5b. The use of platelet components matched for donor HLA and platelet-specific antigens provided in general successful transfusion responses, as measured by CCI’s 12-24 hours post-transfusion. On the other hand, four patients receiving platelets matched only for HLA antigens exhibited poor transfusion responses. In total, 67 transfusions were provided to the six patients. The clinical objective was to maintain platelet counts above 15-20 x 109/L. The percent of transfusion days with platelet counts below 10 x 109/L were as follows: unmatched platelets, 21%; HLA-matched platelets, 33%; and, HLA- and platelet-specific antigen matched platelets, 12%. There were related useful observations from this study. It was determined that HLA and platelet-specific antibodies disappeared during the analysis period in some patients; a number of previous studies have also documented the disappearance of antibodies. Relating to the threshold level for transfusions, minor bleeding was observed when the platelet count was below 5 x 109/L. When the platelet count was between 5-20 x 109/L, the percentage of days with minor bleeding varied between 0% and 22% per patient. Three patients exhibited minor bleeding with platelet counts between 23- 93 x 109/L. Use of ABO-Identical Platelets The advantage of using ABO-matched platelets has been debated for many years. In the late 1980's, three groups, including Heal and Blumberg (20), published data showing that ABO mismatching caused more refractoriness. Heal and Blumberg have championed the use of ABO-identical platelets for patients requiring long-term platelet support. In a prospective controlled trial, patients receiving ABO-identical platelets needed only about half as many transfusion during the first 30 days of platelet therapy or during the first admission compared to patients receiving ABO-unmatched platelets. There was less refractoriness in the group receiving ABO-identical platelets (36% versus 75% of the patients). It has also been suggested from studies with a small cohort of leukemia patients that the use of ABO-unmatched platelet transfusion may result in shorter remission and survival periods (21). The ABO-unmatched platelets may be detrimental because their transfusion leads to increased levels of circulating immune complexes. Practices relating to the use of ABO-matched platelets continue to vary. Many hospitals transfuse ABO-matched platelets whenever possible. Costs of Transfusing Platelets Reports in the literature pertaining to the cost of platelet transfusion are limited. However, it appears that this is an issue that is drawing increasing attention. Petz and co-workers analyzed the use and cost of platelet transfusions at the UCLA Medical Center for non-refractory and refractory hematology/oncology patients. They reported, at the 1997 AABB Meeting, that refractory patients (33% of hematology/oncology admissions) used 71% of the platelet units that were transfused by the hematology/oncology service (3, 22). The total cost incurred by the hospital for platelet transfusions per admission averaged $10,775 for refractory patients and $2,142 for non-refractory patients. The data were stratified for allo-transplants, auto-transplants and other categories of hematology/oncology patients with the cost always being higher for refractory patients. For allo-transplant patients, for example, the average cost per admission for refractory and non-refractory patients was $14,647 and $5,218, respectively. Most of the costs were ascribed to the components that were transfused. In addition, costs incurred for managing refractory patients, including those associated with HLA antibody screening, platelet transfusion monitoring and the coordination needed to provide special products, were considered when deriving the cost per admission. Petz and co-workers also evaluated two methods for evaluating the costs associated with in-patient platelet transfusions, the use of cost/charge ratio methodology and a time-motion procedure which measured the direct costs of the production and administration of platelet transfusions. It was concluded that the use of cost data derived from hospital charges significantly underestimates the true cost of platelet transfusions. Blumberg and co-workers performed a cost analysis pertaining to the use of ABO-identical platelets that were also leukocyte depleted (23). The use of such products was associated with substantial cost savings. Much of the decrease in cost reflected the reduced use of platelets. Hospital charges for various services were also less for patients receiving filtered ABO-identical platelets compared to unfiltered ABO-unmatched platelets. Recent developments Research on alternative ways to maintain platelet-related hemostasis is continuing. Infusible human platelet membrane preparations have been developed as a therapeutic treatment by PRP, Inc. Currently, Cypress Bioscience, Inc. is evaluating this product in human trials. It was reported at the 1997 ASH Meeting that in a phase I trial, these membrane preparations controlled bleeding in 17/26 thrombocytopenic patients, including some (7/12) who were refractory to standard platelet transfusions (24). As a follow-up, a phase II trial with refractory patients is being conducted at ten medical centers in the United States. A dose-ranging study is planned for 1999. Thrombopoeitin and a truncated version, pegylated recombinant human megakaryocyte growth and development factor (PEG-rHuMGDF), are being evaluated as a means for reducing the need for platelet transfusions, especially following stem cell transplantations. MGDF, developed by Amgen, has been shown to have a favorable safety and pharmacodynamic profile (25). However, overall its effect on reducing the need for platelets has been limited. Amgen is also evaluating how MGDF could be used to increase the platelet levels of normal individuals prior to apheresis donation. References
Last modified: |
