Today the military's use of Factor VII has been significantly scaled back. Physicians at the Baghdad hospital said last summer that they recalled using it about a dozen times in the first six months of 2008. During a trip to the same hospital two years previously, The Sun twice saw a dozen doses given in a single day.

Finding: Transfusions of fresh whole blood, a rarity in civilian medicine, became a standard treatment early in the war, based mostly on anecdotes and theoretical arguments. But the practice unwittingly exposed 20 or more patients to hepatitis, studies of its effectiveness have found mixed results, and it is now used only in emergencies.

The first guideline for Factor VII called for giving two units of fresh whole blood along with the drug.

The concept goes against standard practice in the United States, which is to break down blood into components - platelets, plasma, red blood cells - and give only the components a doctor thinks are required. Components broken down in a laboratory are rigorously tested for diseases such as hepatitis and HIV, whereas whole blood can only get cursory testing, if that.

During the run into Baghdad in 2003, fresh blood was often the only thing available, particularly as a source of platelets, which don't travel well but are important for stopping blood loss. In interviews, Army doctors recounted success stories of the "walking blood banks" early in the war.

But even as medical facilities in Iraq improved and the emergency need for whole blood diminished, many military doctors kept using it, based solely on their observations that fresh whole blood worked better than frozen components. According to one Army study, more than 6,000 units of whole blood were transfused by military doctors in the first four years of the war - much of it because no other blood was available but also because some doctors preferred it.

For months a debate raged between doctors in the United States and their counterparts in Iraq, according to e-mail traffic shown to The Sun. The Navy's top trauma adviser called whole blood "like nothing I have ever seen" and said he orders it for every massive transfusion patient. The chief of transfusion research at Walter Reed Army Medical Center called its routine use "indefensible" and said: "You are risking death, malignancy or lifelong infection with hepatitis."

Schoomaker said in December that 20 units of whole blood contaminated with hepatitis made it through the military's screening procedures in the war zones. At least one soldier contracted Hepatitis C from a transfusion and is now in treatment. While the odds of infection are still low, given the thousands of units of blood transfused during the war, the military issued a guideline in early 2007 limiting the use of fresh blood to times when nothing else is available.

But the debate continued. Last summer, Army surgeons in Iraq fought to save a 32-year-old Army captain who kept bleeding from multiple gunshot wounds even after numerous surgeries, 50 units of transfused components and a dose of Factor VII. Desperate, they called Capt. Victoria McCarthy, officer in charge of the blood bank, and asked about activating a whole blood drive.

"I said no," McCarthy said. "I had components available, and told them they couldn't use soldiers as guinea pigs."

The captain bled to death well before any whole blood would have been available.

"There's no science to suggest that fresh whole blood leads to better outcomes, just some idiosyncratic views about health and disease that are based on anecdotes," said Dr. David Walker, chairman of the Defense Health Board's subcommittee on emergency transfusions, which issued a report last year calling whole blood "undesirable." "If data comes out that it's better, that's different."

Controversy over the experimental nature of Army medical care continues in Iraq over a procedure that Army leaders consider "the single most important advance in trauma care for hospitalized civilian and military casualties from this war."

The standard procedure for an emergency blood transfusion before the war started was to give a patient red blood cells, the basic oxygen-carrying component in real blood, and then order other components after diagnosis.

But beginning in 2002 in Afghanistan, and continuing in 2003 in Iraq, reports from battlefield surgeons suggested that patients were bleeding to death because their blood had become too diluted with red cells or fluids to clot, not simply because of their injuries. In 2004, doctors in Iraq recommended thawing frozen plasma - a vital component in the clotting process - and using it at the beginning of the transfusion process. Each time a casualty got a unit of red cells, the guideline stated, he should also get a unit of plasma.

The concept behind the 1:1 ratio was simple: It's the same ratio found in whole blood. The new procedure was institutionalized by an Army-wide memo issued in January 2007 and became known as "damage control resuscitation."

But plasma has been implicated in lung infections and other complications, and the new ratio quadrupled the standard dose of plasma. Like some other guidelines, the 1:1 ratio was initially based solely on anecdotes, theories and computer modeling performed in San Antonio.

In 2007, in a scientific paper based on the treatment of patients in Iraq, Army researchers concluded that increased use of plasma improved survival by as much as 60 percent. At the time, the Army said it was the only published research based on the treatment of human patients.

Damage control resuscitation has since been widely adopted by civilian trauma centers. "It's a paradigm shift," said Dr. Stephen Smith, who implemented the practice at the Via Christi Regional Medical Center in Wichita, Kan., before moving last year to the new Virginia Tech medical school.

Said Dr. Juan C. Duchesne, a trauma surgeon and researcher at Tulane University Hospital in New Orleans: "It turns out we've been doing it wrong for the last 50 years."