What were they thinking?

Early last year, a task force of thoracic surgeons and cardiovascular anesthesiologists met to write new clinical practice guidelines for how to conserve blood and prevent blood transfusions during heart surgery. Just a few months earlier, the FDA had learned that Bayer, the maker of Trasylol (aprotinin), a blood-clotting drug used in surgery, had hidden results of its own study of the drug from the FDA and from an FDA advisory committee. The story made national headlines.

The hidden data showed that patients who received aprotinin were at increased risk for death, kidney failure, congestive heart failure and stroke compared to patients who received other anti-bleeding drugs. Yet here is what the guideline, published in The Annals of Thoracic Surgery in May 2007, recommended:

1. High-dose aprotinin to reduce blood transfusions, blood loss and the need for followup surgery to stop internal bleeding in high-risk heart surgery patients. (The guideline notes that the benefits of use should be balanced against the increased risk of kidney dysfunction.)

2. Low-dose aprotinin to reduce blood transfusions and blood loss in other patients having heart surgery.

3. Lysine analogues to reduce blood transfusions and blood loss in patients having heart surgery. The guideline states, however, that the safety of lysine analogues "is less well studied compared with aprotinin."

The recommendations are labeled "Class I" (the classification for treatments for which the benefit clearly outweighs the risk) and "level of evidence A" (meaning sufficient evidence is available from randomized trials or meta-analyses). The authors rejected the safety concerns with aprotinin that had been raised by Dennis Mangano and colleagues in a 2006 study published in the New England Journal of Medicine.

A subsequent study published by Mangano's group in the Journal of the American Medical Association in February 2007 was not discussed. It is unclear whether the JAMA study, which found that aprotinin raises the risk of death over five years after surgery, was available at the time the guideline was being written. If not, one wonders why such important information did not merit an update or revision.

The votes on the three recommendations were 15 to 2, 15 to 2, and 16 to 1. Eight out of 17 of the guideline authors disclosed ties to Bayer in the form of lecture and consulting fees and/or research support. Would the outcomes of these votes been different if so many of the authors had not had ties to the maker of aprotinin? If the guidelines had not endorsed continued use of aprotinin, could lives have been saved?

On November 5, 2007, the FDA requested a marketing suspension of aprotinin in response to preliminary results of a trial showing an increased risk of death in patients receiving aprotinin. A week later -- more than nine months after the JAMA study, the Society of Thoracic Surgeons finally announced that the portion of the guideline relating to use of aprotinin was under review.

--PM