Sanne de Bruin

117 InPUT- pilot study in the Netherlands Discussion In this pilot study, we observed that patients in our ICU were relatively restrictively transfused, and that pre- and post RBC transfusion Hb levels were similar in bleeding and non-bleeding patients. Furthermore, we showed the feasibility of our study protocol for an international point prevalence study of transfusion practice in ICUs. This study serves as a pilot study which is ready to be enrolled internationally as an observational cohort study. In this small population the reported pre-RBC transfusion Hb levels of 7.6 (6.7-7.7) g/ dL were lower than a recently published large observational cohort study in the ICU where a mean nadir Hb level on day of transfusion of 8.3 g/dL was reported 2 . Also the proportion of patients transfused with RBCs (12.5%) was lower than in this study where 26.3% received one or more RBC transfusions 2 . This lower proportion could be part of an ongoing downward trend in RBC transfusion, since the proportion of patients who received an RBC transfusion during ICU admittance has been declining the last decade. In 2002 37.0% of ICU patients was transfused 9 , while in 2018 the same research group reported that only 26.3% received RBCs during their ICU stay 2 . Another explanation of the differences with current literature is the large proportion of post cardiothoracic surgery (35%) patients in this patient cohort. These patients were admitted with a relatively high haemoglobin level of 11.8 (10.5-12.1) g/dL and were ad- mitted for a median length of ICU stay of two days, resulting in only one patient in this subpopulation receiving RBC concentrates. Also, the proportion of patients receiving platelet concentrates in our study was only 6.3% of the patients during ICU admittance, which is slightly less than observed in the United Kingdom in 2012 where 9.0% of a national patient cohort received platelets during ICU stay 3 . Also, in that study more than 40% of the patients who received plate- lets had a platelet count ≥ 50 x 10 9 cells/L 3 , while in our study, only 18% of the patients who received platelets had a platelet count ≥ 50 x 10 9 cells/L. In addition, all non-bleed- ing patients who received platelet concentrates prophylactically prior an invasive pro- cedure or without any upcoming procedure had a platelet count of ≤ 50 x 10 9 cells/L and ≤ 10 x 10 9 cells/L, respectively. Only bleeding patients had a platelet count ≥ 50 x 10 9 cells/L prior to transfusion. This is in accordance to our local transfusion guideline. Unnecessary platelet transfusions should be avoided, especially since multiple RCTs in different patient populations have shown an increased mortality in patients transfused with a liberal platelet transfusion strategy 19,20 . 4