The Long and Successful Journey of Prophylactic Platelet Transfusion

It has been standard policy for the past 40 years to administer platelet transfusion prophylactically to prevent hemorrhage in patients receiving intensive therapy for acute leukemia and in recipients of stem cell transplantation. And this approach has worked! Despite prolonged, severe thrombocytopenia often in the settings of mucositis and severe infections, fatal or life-threatening bleeding events are uncommon. Indeed, contemporary clinical trials in older patients receiving intensive induction chemotherapy report overall early mortality rates of 5 to 10 percent, with fatal bleeding of approximately 2 percent.¹  Furthermore, advances in platelet collection and storage²  and the near-universal application of pre-storage leukodepletion has markedly reduced the incidence of transfusion reactions,³  alloimmunization, and refractoriness to transfusion.⁴ 

I first became involved in this area of clinical research in 1972 when the gospel was to transfuse prophylactically at platelet counts less than 20,000/µL, likely based on a 1962 paper from the National Cancer Institute by Drs. Lawrence A. Gaydos and Emil J. Freireich demonstrating very high rates of bleeding in children with leukemia, explained with a rather primitive graph, suggesting that the rates increased further at counts less than 20,000/µL.⁵  It is frequently instructive to go back and read what was actually written (i.e., how the authors interpreted their own data). In their prescient words: "No ‘threshold’ platelet level was observed. A ‘threshold’ concept would be useful for the description of a direct causal relationship. For platelet values above the threshold would be adequate to prevent hemorrhage, whereas at levels below this threshold, hemorrhage would occur. This type of relationship does not exist (italics added) between platelet count and hemorrhage… More proximate causes interact with platelets to produce hemorrhage. It is likely that anything done to maintain or raise the platelet count would decrease the frequency of hemorrhage." In other words, wouldn't it be nice if we had platelet transfusions!

The clinical realities were very different then: more gram-negative bacteremias, lousy and more toxic antibiotics, inferior chemotherapy with longer periods of aplasia, poor venous access, the use of aspirin as an antipyretic, and of course, the absence of platelet transfusion. Indeed, as late as 1987, a National Institutes of Health Consensus Conference on platelet transfusion⁶  concluded that, "the patient with severe thrombocytopenia may benefit from platelet transfusion… It is common practice to use a preselected level of thrombocytopenia to decide when to transfuse platelets prophylactically. The figure of 20,000/ mm³ is often used. This figure is based on older studies with potential defects… Recent evidence suggests that this number might be lower for some patients based on clinical judgment and close observation." [Italics added.]

This rather conservative early foray into "evidence-based medicine" contained enough qualifiers and ambiguities to gratify an army of plaintiffs’ attorneys.

Subsequent clinical trials fortunately provided further clarification, with two randomized studies in patients with leukemia showing that the rates of hemorrhage were similar using a transfusion threshold less than 10,000/µL compared to the less than 20,000/µL threshold group, with a roughly 20 percent reduction in the number of platelet transfusions administered.^7,8  Importantly, the rates of serious hemorrhage were very low in these trials, with only one episode of fatal bleeding. Other observational studies suggested that even lower thresholds can be considered by experienced clinicians in a subspecialty unit.⁹ 

More recently, in a randomized study conducted largely in the United Kingdom, the new "standard" prophylactic threshold of less than 10,000/µL was compared to a sort of "therapeutic" approach in which transfusions were administered only if patients were actually experiencing mild bleeding.¹⁰  A heterogeneous population of patients, 70 percent of whom had undergone autologous stem cell transplantation (ASCT), was studied, and the authors came down in favor of the prophylactic approach, concluding: "Therefore, this study did not show that a no-prophylaxis strategy for platelet transfusions was noninferior to a prophylaxis strategy in relation to the frequency of bleeding events of WHO grade 2, 3, or 4." A lot of double negatives, but critically, there was only a 1 to 2 percent incidence of grades 3/4 bleeding with the major "difference" being related to a 5 to 6 percent difference in the rate of more minor grade 2 bleeding in favor of the prophylaxis arm. Approximately 40 percent of ASCT recipients required no platelet transfusions, presumably due to both normal bone marrow function and platelet counts at the time of transplantation, as well as rapid reconstitution using peripheral blood as the source of stem cells.

Very similar results were reported from a large randomized trial from Germany with their conclusions leaning toward the no-prophylaxis approach, except in patients with acute leukemia.¹¹  Indeed, in a subsequent publication reanalyzing essentially the same data, the UK group modified their conclusions somewhat, acknowledging that different approaches might be appropriate in different clinical scenarios.^12,13  Results from both studies reinforce the observations that with proper management, severe hemorrhage is rare even in critically ill severely thrombocytopenic patients. Major kudos to our microvascular system (!!), with additional credit due to increased effectiveness of the blood supply system and more knowledgeable clinicians. And, as suggested in 1962, it is not just the number of platelets. Transfusions should be considered at higher counts for those with sepsis, hyperleukocytosis, or coagulation abnormalities as in patients with acute progranulocytic leukemia, and are not mandatory, even at lower counts, in clinically stable, nonbleeding patients with myelodysplasia or aplastic anemia and chronic thrombocytopenia.

There remain some important issues surrounding prophylactic transfusions that require further attention. The most troublesome section in the American Society of Clinical Oncology platelet transfusion guidelines relates to the need for specific thresholds prior to invasive procedures.¹⁴  Superstition and dataless-driven guidelines from other subspecialty groups abound, with most hovering at levels of approximately 50,000/µL, often referring to older observational studies.^15,16  We have all had patients returned from invasive radiology because the platelet count was 47,000/µL with instructions to send them back when it's "higher than 50,000." Randomized data are not and never will be available because fortunately, the "background" rates of misadventures are so low, and sample sizes would be prohibitive. There have been reasonably large descriptive studies in patients undergoing central line placement, lumbar puncture, and bronchoscopy with bronchoalveolar lavage documenting safety at counts in the range of around 20,000/µL and likely even lower.^17-20  More such observational studies in patients undergoing surgery and invasive interventions should be strongly encouraged with a major award given to whoever persuades those who wiggle the needles and knives to adopt a more sensible approach.

Nonetheless, current approaches have largely eliminated serious hemorrhage, which has in turn contributed substantially to the higher cure rates now enjoyed by patients with aggressive hematologic malignancies.