What is clinically significant bleeding?

Abstract

Bleeding is an important cause of morbidity and mortality in patients with thrombocytopenia from a range of conditions, including bone marrow failure. In this setting, serious bleeding and its consequences, or the potential for such bleeding, is feared by patients, families, and clinicians alike and drives much of our current practice. Platelet transfusions are often administered at predefined thresholds in an effort to prevent bleeding or as therapy when bleeding occurs. Patients with hematologic malignancies frequently develop profound thrombocytopenia due to their underlying disease or treatment. They also account for the majority of all platelet transfusions and a substantial proportion of all red blood cell (RBC) transfusions given around the world. However, many patients continue to experience bleeding despite widespread use of prophylactic platelet transfusions, highlighting the need to identify those most likely to develop problematic bleeding, and the importance of risk-adapted management approaches, including alternatives to platelet transfusions. Furthermore, the definition of a “clinically significant” bleeding episode remains to be established and validated, and this is exacerbated by the fact that few data are available on whether the presence of “minor” bleeding heralds more serious events. The original World Health Organization (WHO) classification assigned a range of bleeding signs and symptoms and their consequences to grades from 0 (no bleeding) to 4 (debilitating or fatal bleeding). Clinically significant bleeding outcomes in trials are now most commonly described using various modifications of the original WHO scale, typically including Grade 2 (mild blood loss) or higher bleeding events. Grade 1 and 2 bleeding events, which account for the vast majority of reports, are clinically heterogeneous, and how they are defined differs widely among trials. Are they all “clinically significant”? This may depend on who is asking, and answering, the question, as well as the population involved, how the bleeding is classified, and the intended use of the information. For example, some Grade 1 and 2 bleeds may cause distress because they look bad (eg, subconjunctival hemorrhage) or feel bad (eg, epistaxis). These sorts of bleeds may worry the patient, family, or clinicians (who have historically been more concerned about the implications of “wet purpura” such as blood blisters in the mouth than some other types of bleeding)—but they are not inherently “serious” in that they do not require specific intervention. These bleeding episodes may be clinically significant if they predict the occurrence of subsequent worse bleeding events and outcomes. However, we do not know which (if any) types of Grade 1 or 2 bleeding predict more severe bleeding, how, or why. Balitsky and colleagues present a detailed analysis of bleeding assessments from three randomized controlled trials (RCTs) to describe the frequency of the different signs and symptoms of WHO Grade 1 and 2 bleeding events, and to examine associations with more severe (Grade 3 and 4) bleeding. The data had been prospectively collected during a trial of RBC transfusion thresholds, the Strategies for Transfusion of Platelets (SToP) trial, and the Pathogen Reduction Evaluation and Predictive Analytical Rating Score (PREPAReS) trial, and comprised 315 adults with thrombocytopenia secondary to hematologic malignancy or chemotherapy who had sufficient baseline data and at least one daily bleeding assessment. The majority (>90%) of patients were being treated for acute leukemia, mostly acute myeloid leukemia. For all three RCTs, bleeding assessments were performed daily by trained research staff, plus, in some cases, by patient self-report. The bleeding assessment tools, data items captured in case report forms, and modified WHO classifications were slightly different for each trial but were sufficiently similar to allow pooling of the data. Transfusion guidelines and thresholds were also similar. From a total of 5476 available bleeding assessments (median, 17 [interquartile range, 12-22] per patient), bleeding of Grades 1-4 was documented in 61.8% of daily reports. Where bleeding was recorded, 57% of the episodes were assessed as Grade 1, 41.3% as Grade 2, 0.9% as Grade 3 and 0.8% as Grade 4. The Grade 1 and 2 events were further “deconstructed” from the reports into 20 individual types of bleeding signs and symptoms. The majority were skin-related (ecchymosis, petechiae, and purpura). Hematuria (macroscopic or microscopic) was reported in 132 (4%) Grade 1 or 2 assessments and was the only sign or symptom with a statistically significant association with more severe Grade 3 or 4 bleeding episodes (hazard ratio, 16.1; 95% confidence interval, 4.4-59.2; P < .0001). Of the laboratory results examined, Received: 9 January 2021 Accepted: 9 January 2021