Epidemiology of platelet transfusion in critically ill patients: a multicentre observational prospective study

Background Platelet transfusion requirements in critically ill patients is highly variable. Describing platelet transfusion epidemiology and adherence to guidelines might help to determine how to improve transfusion practices. Study design and methods This is a pre-planned secondary analysis of a prospective, multicentre, observational study analysing platelet transfusion response. The present study aims to describe platelet transfusion indications, platelet count prior to transfusion and adherence to guidelines. Results A total of 975 platelet transfusions were administered to 310 patients. Prophylactic transfusions given to thrombocytopenic patients and to prevent bleeding prior to an invasive procedure accounted for 41.7%; and 19.7%, respectively; while 210 (21.5%) and 166 (17.0%) platelet concentrates were given to treat major and minor bleeding, respectively. Median platelet counts were 14 × 10 9 /L (Interquartile range (IQR) 8–23 × 10 9 /L) when considering prophylactic transfusion, 21 × 10 9 /L (IQR 12–36 × 10 9 /L) prior to invasive procedure and 43 × 10 9 /L [IQR 27–66 × 10 9 /L] prior to therapeutic transfusions. Fifty eight percent [95% confidence interval (CI) 55–61] of platelet transfusions were given in accordance with guidelines. In the exploratory multivariable analysis, the model-based predicted probability of non-adherence was 76.4% for prophylactic platelet transfusions compared to 28.5% for transfusions given prior to an invasive procedure or to treat bleeding, corresponding to an absolute difference of 47.8 percentage points (95% CI 40.7–55.0). Conclusion In this large study, 42% of platelet transfusions were not aligned with guidelines. Prophylactic platelet transfusion was the indication with the lowest adherence to guidelines. These results emphasise the need to better define platelet transfusion indications.