My first exposure to the literature on transfusion thresholds in critically ill patients was through the landmark study by Hebert et al. , which compared restrictive (7 g/dL) versus liberal (9 g/dL) transfusion strategies. The authors concluded that a restrictive approach was at least as effective as, if not superior to, the liberal strategy.
My second encounter with this topic came from the guideline published by West et al. , which reinforced a similar philosophy toward transfusion practices. Since then, I have often referred back to another milestone study in the New England Journal of Medicine, which examined transfusion thresholds not only in patients requiring transfusion but also in those with significant comorbidities, such as cirrhosis. While the evidence has long supported slightly higher thresholds for certain populations, particularly cardiac patients (around 8 g/dL), a recent study in the Annals of Internal Medicine (2025) has shifted this discussion by proposing a new threshold of 10 g/dL in acute myocardial infarction, albeit supported by low-level evidence.