Sanne de Bruin

118 Chapter 4 While plasma transfusion is possibly beneficial in bleeding patients, plasma transfusion to correct prolonged coagulation time in the absence of bleeding is not recommend- ed 14,15 . This is in accordance with our findings that only patients who were actively bleed- ing received plasma, but in contrast with a large study in the UK where only a third of the plasma transfusions was given to treat bleeding 21 . Also the proportion of patients of this cohort receiving plasma was two times higher (12.7%) than in our study 4 . The correction of the PT in our patient population was minimal after plasma transfusion. In this study tranexamic acid was administered five times. Multiple studies showed a re- duced need for transfusion during surgery when inhibiting fibrinolysis using tranexamic acid 22 . Also in trauma patients a decreased mortality was found when administering tranexamic acid early after trauma 23 . However, it is unclear how we can extrapolate these results to bleeding patients admitted to the ICU. To our knowledge, RCTs in this patient population comparing tranexamic acid to placebo have never been performed. Themajor limitation of this study is the small number of included patients. With only 48 patients in this pilot study, the transfusion numbers are too dependent on a few specific patients; two patients were responsible for 30 transfusion events (79%). Furthermore, this was a single centre study, thus limiting generalizability of the findings. A large international study is needed to draw valid conclusions on transfusion prac- tices. Ideally, different subpopulations need to be included such as patients with and without bleeding and patients with acute coronary syndrome, brain injury and patients on extracorporeal membrane oxygenation. Current rationale is that these patient cat- egories might need higher Hb levels, while evidence for this is limited. This protocol enables the researchers, when included a sufficient number of patients, to compare transfused and non-transfused patients and examine differences inwhy certain patients are transfused differently while having similar Hb levels, cell counts and coagulation tests. Numerous studies have studied RBC transfusions in large ICU cohorts, however, international studies examining plasma transfusion, platelet transfusion, and admin- istration of coagulation factors are limited. The feasibility of this protocol was shown by the acceptable amount of time per patient and number of registered blood product administration. However, based on this pilot study, the study protocol of the international observational cohort study was adjusted. Time consuming data collection that did not result in reliable and useful informationwas removed from the study protocol, including daily fluid balance and additional questions for the non-transfused patients. Also the workflow of the data collection was improved