The results were mixed. The mortality rate was an encouragingly low 3.2%, but Carson also discovered that the risk of complications or death was higher in people with heart disease. "In some circumstances," he says, "blood is lifesaving. When people get very low blood levels, their risk of running into trouble is substantial, and if you're old or have cardiovascular disease, that risk may be even greater. So I recommend caution."

Shander is not put off by such fears. "Medicine is very conservative," he says, "which can be good, since it protects doctors against going along with every unproved technique that comes along. But it's imperative that we develop a mind-set where we look at refusing blood not as an obstacle but as a challenge."

One way of responding to the challenge might well be the development of artificial blood. The quest for a blood substitute reaches back to the 17th century, when scientists tried to transfuse animal blood and other products into humans. Several blood substitutes are undergoing clinical trials in the U.S. and Europe, and one, which seems to carry oxygen like its genuine counterpart, has been tested successfully in heart-surgery patients in Europe.

But for Shander, the problem is a more personal one. "When we're challenged, we extend ourselves," he says. "Some of my colleagues have adopted bloodless medicine purely as a technique. Others have learned that it also has an impact on ethical and humanistic values. I feel that once you become philosophically committed to practicing bloodless surgery, the benefits to patient and physician alike become more and more apparent. Those are my greatest rewards."