The authors sought to examine the relationship between transfusion and in-hospital mortality in patients with acute myocardial infarction (AMI). For this they interrogated the Cerner Health Facts database that included close to 35,000 AMI patients in 57 hospitals for the years 2000 to 2008. A total of 1,778 patients received at least 1 transfusion.
Their most striking finding was that, when attempting to match transfused with non-transfused patients, the vast majority of patients had non-overlapping propensity scores, indicating that transfused and non-transfused patients have markedly different clinical profiles. In other words, it appears that the physicians’ clinical judgment leading to transfusion is governed by factors that cannot be captured by databases. Consequently, transfused and non-transfused patients are fundamentally different patient populations that cannot be compared.
This important observational study casts serious doubts on the validity of the conclusions of previously published observational studies using less robust statistical methods. As concluded by the authors and several others before them, we definitely need randomised trials to establish the benefits and risks of transfusion during AMI. Fortunately, it appears that several such studies are underway or in planning (2 pilot studies have already been published).
– Jean-François Hardy
Discuss this article on the Nataonline forum