"There [have] been four big randomized trials and they all four have shown no difference between overall outcome between thresholds of 8 and 10," Frank said. "Yet even though they show no difference, people in medicine are still using 10 as the threshold, even though 8 is just as good."
The biggest study to date was published in December by the FOCUS research group (Functional Outcomes in Cardiovascular Patients Undergoing Surgical Hip Fracture Repair), Frank explained. "Those researchers found that by lowering hemoglobin threshold to 8 (instead of) 10, they used 66% fewer blood transfusions and they had no different outcomes between the two groups in length of stay, heart attack, stroke, death, and even the ability to walk," Frank said.
"Their conclusion was that if there's no benefit from giving extra blood, then all (that's left is the additional) cost and the potential for harm by giving more blood than is needed."
In their latest paper in Anesthesiology, Frank and his co-authors quantified the transfusion practices for each of 44 Johns Hopkins surgeons, and then compared their use of blood by type of patient and type of procedure. For example, they found that among four surgeons performing the Whipple procedure, there was a 1.8 g/dl difference between the point at which a transfusion was initiated, and a 1.9 g/dl difference between the highest and lowest target levels at which transfusions were stopped.
Other procedures in which there was great variation in either the hemoglobin "trigger" and/or the hemoglobin "target" were posterior lumbar fusion, and primary coronary artery bypass grafts.