Sanne de Bruin

37 Transfusion practice in the non-bleeding critically ill; an international online survey – The TRACE Survey The finding of this survey that 57% and 43% of physicians would consider to transfuse plasma to correct the prolonged INR in patients who used vitamin K antagonists prior an invasive procedure or without a planned invasive procedure, respectively, is striking in the absence of evidence for this practice. To our knowledge, the influence of base specialty of intensivists on transfusion practice has not been studied before. Our survey showed that doctors with a base specialty anaesthesiology transfuse more liberal than those with internal medicine as base spe- cialty. It might be that doctors with an internal medicine background are more aware of the harmful side effects of blood products, alternatively, anesthesiologists may tend to treat patients at higher risk of bleeding, and these practices spill over into the ICU. The strength of this survey is the large number of respondents. However, both the anonymous character of this survey and the origin of the respondents might have in- troduced a selection bias and limits therefore the worldwide generalizability of our findings. Theoretically, it is possible that multiple respondents are employed in the same hospital, however also within hospitals heterogeneity in transfusion practice may exist. Furthermore, the number of respondents who did receive this survey but did not fill it in is unknown. We cannot exclude that non-responders transfuse differently than the responders of this survey. It may be possible that physicians with more interest in transfusion practice and thus with more awareness of the possible side effects of transfusion are over presented in this survey. In addition, due to the study design it was not appropriate to perform multivariable analysis. As result, it was not possible to exclude the presence of confounding variables on the observed significant associa- tions. And finally, as with any clinical practice survey, the reported transfusion practices might differ from actual transfusion practices. Ideally, these results are confirmed in a prospective cohort study. Conclusion In conclusion, in the general non-bleeding ICU population, the reported RCC transfusion practicewas rather restrictive, however, in certain subpopulations including the critically ill with septic shock, higher applied Hb thresholds were reported, which deviates from current evidence. For other subpopulations such as patients with ARDS and patients on ECMO well powered RCTs are needed. In addition, optimal platelet thresholds are cur- rently controversial and more awareness is necessary for correct indications of plasma 2