Feb 222012

There's not much data to guide the transfusion of red blood cells. In 1999, there was the beautifully executed and practice-changing Transfusion Requirements in Critical Care (TRICC) trial, which showed that a restrictive transfusion strategy (hemoglobin trigger of 7 g/dL) in ICU patients resulted in a non-stat.significantly lower mortality (19% vs. 23%) compared to a liberal transfusion strategy (Hb trigger of 10 g/dL). But otherwise, most trials have been small and insufficiently convincing to herd us physicians into a single practice pattern.

One tricky thing about TRICC was that although patients with coronary artery disease were allowed into the trial, and subgroup analysis suggested they also did as well or better with restrictive transfusions (as long as they were not in the midst of a coronary event), they could be excluded from trial enrollment by their physicians on an ad-hoc basis, and so the applicability of TRICC to those with coronary artery disease has always been subject to debate.

Enter Jeffrey Carson, Michael Terrin, Jay Magaziner et al (the FOCUS investigators).

Between 2004-2009 at 47 sites in the U.S. and Canada, they began by enrolling people aged 50+ undergoing hip fracture repair who had a hemoglobin < 10 g/dL within 3 days after the surgery, and a history of cardiovascular disease (coronary artery disease; myocardial infarction; congestive heart failure; stroke/TIA; ECG with Q waves; or peripheral vascular disease). They got about 1,200 of these folks. However, 18 months in, it became clear they weren't going to get enough subjects to achieve power to detect a 7% absolute difference between groups. They then expanded the criteria to include anyone with cardiovascular risk factors, and anyone with chronic kidney disease too. With this boost of 800 people, they got to n=2,016. Median age was 81 years old; 75% were women.

They randomized them to either a restrictive (hemoglobin > 8 g/dL) or liberal transfusion strategy (hemoglobin > 10). Blood could also be given for any concern for active cardiac pathology. The primary outcome was a composite of death or extreme disability (the inability to walk 10 feet without human help 60 days after surgery).


  • No difference between groups in the primary outcome (death/inability to walk at 60 days). Both were 35%.
  • No difference in numerous prespecified secondary outcomes (myocardial infarction, unstable angina, death at 30 days, etc., etc).
  • The restricted-group got far less blood. 59% got no blood after surgery, and only 37% got 1 or 2 units; only 3% of the liberal-group got no blood, and 75% got at least 2 units. Overall, the liberal group got 1,866 units, compared to 652 for the restricted group.
  • Troponin levels were collected routinely at days 1 and 4; there was no difference between groups (6% in each had an isolated elevation, without acute coronary syndrome).
  • Nitpick: The restricted group had slightly more blood loss during surgery, and got slightly more blood before randomization (80 units total, compared to the 1,200 unit difference between groups after randomization).
  • Subgroup analysis revealed that men in the liberal transfusion group did worse, with an odds ratio of 1.45 for death/disability; this wasn't a pre-specified analysis and may have been due to chance.

In the end, their 95% confidence interval allowed for the possibility of, at most, a 3.7% increase in death/disability due to the restrictive strategy (or an 8% decrease in that outcome).

Clinical Takeaway: This well-done randomized trial adds to the growing weight of evidence that liberal blood transfusion does not seem to improve outcomes, and may harm some patients. Although only ~60% of subjects had documented cardiovascular disease, the absence of an excess of cardiovascular events (or any negative outcome) in the restrictive transfusion group should provide physicians confidence that a conservative approach to blood transfusions (hemoglobin trigger of 8 g/dL), carefully monitored, is appropriate for most patients with stable cardiovascular disease or risk factors undergoing major surgery. With more than half of the patients aged 80 or older, this trial should also dissuade physicians who believe old age alone is a reason to transfuse red blood cells for postoperative anemia.

If you haven't yet, check out Tim Hannon's blog The Bloody Truth for a fresh take on the most current news in blood transfusions and blood banking.

Carson JL et al (the FOCUS investigators). Liberal or Restrictive Transfusion in High-Risk Patients after Hip Surgery. N Engl J Med ePub December 14, 2011.

Get our weekly email update, and explore our library of practice updates and review articles.

PulmCCM is an independent publication not affiliated with or endorsed by any organization, society or journal referenced on the website. (Terms of Use | Privacy Policy)


Restrictive blood transfusion was fine for high-CV-risk patients after hip fracture repair (RCT, NEJM)