Roughly 3 1/2 years after the war in Iraq began, the Department of Defense is only now beginning to implement a comprehensive computer system to record the treatments and outcomes for the 21,500 casualties from Iraq, about 6,500 of whom had injuries serious enough to require evacuation. Researchers are also trying to create a historical database using paper medical records, but it is a tedious process, and doctors say the accuracy and completeness of paperwork generated in a war zone are suspect.

As such, the military has no authoritative method of determining how many patients have received Factor VII and how many subsequently developed blood clots. Nor can it determine the rate of complications among patients who did not get the drug. Holcomb said that the military has administered the drug to more than 1,000 troops, and medical supply records from the Army show that the service has purchased the equivalent of about 2,000 standard doses since early 2004, but the Army doesn't keep track of how much has been administered.

And without computerized data, trends in the treatment of combat casualties are difficult to spot, military doctors say. Many physicians work only in Iraq or Germany for a few months before returning to non-combat jobs in the United States. And doctors outside Iraq say they don't always know which patients received Factor VII inside Iraq because complete medical records may not travel with each casualty from hospital to hospital. Patients occasionally arrive in Balad or Germany with notes from frontline surgeons written on their bandages.

Without hard evidence about Factor VII, military doctors say surgeons in Iraq -- "downrange," in military parlance -- are unlikely to worry about potential complications in light of the anecdotal evidence they see that the drug is working.

"Unless you have a formal policy, it's going to be pretty hard to convince surgeons downrange that they should stop using Factor VII because you're seeing a lot of clots. They'll say, `Well, I'm seeing a lot of bleeding,'" said Dr. William L. Jackson, a former intensive care physician at Walter Reed who left the Army in June to pursue a civilian medical practice.

"I really do think we do a good job by all of these soldiers and that they get excellent medical care. But figuring out something like whether Factor VII is hurting more people than it's helping is not the kind of thing that the Army does particularly well."

Others say that individual experience and anecdotes are shaky grounds on which to base medical decisions.

"To say that because you're in a war everything you do is right suggests to me a level of arrogance that can only lead to a poor outcome," said Dr. Andrew F. Shorr, formerly a pulmonary and critical care specialist at Walter Reed who recently left the Army and took a similar position at the Washington Hospital Center.

"Think about it. If you've got young soldiers having weirdo strokes, and you know they've been exposed to a drug like Factor VII, how long can you presume you don't have a safety issue? Just because you have the best of motives doesn't mean you don't have mediocre methods that are doing more harm than good."

Defense officials denied The Sun's request to review autopsy reports of soldiers killed in Iraq and Afghanistan, first on privacy grounds and then, when the newspaper requested the documents with names or other personal information removed, on the grounds that the reports contain intelligence that could be exploited by the military's enemies.

Three wounded soldiers

The Sun was able to identify a handful of wounded soldiers, either by witnessing their treatment in Iraq or reviewing their medical records weeks later, who were injected with the drug and later suffered unexpected episodes related to blood clots, including stroke, pulmonary embolism and heart attack.

Capt. Shane R. Mahaffee, 36, wounded by a roadside bomb near Hilla, Iraq, on May 5, was injected with repeated doses of the drug in the emergency room and during surgery in Baghdad, and four days later he suffered a pulmonary embolism -- a PE, in medical jargon. He died May 15 of infection and respiratory problems.

"Had he not had that PE, he probably would have survived," said Warren Dorlac, who oversees all the trauma patients treated at Landstuhl. "The PE caused him to get intubated, and then get pneumonia, and then he spiraled downhill."

Pfc. Caleb A. Lufkin, 24, injured by a bomb on May 4 in southern Baghdad and given Factor VII immediately upon arrival at the Baghdad hospital, suffered a blood clot in his lung two weeks later during surgery on his leg. The procedure was stopped, and he was revived and placed on anti-coagulant drugs.

Lufkin died a week later, under similar circumstances, during a surgery that would not have been necessary if the blood clot hadn't stopped his earlier operation. His autopsy report, obtained from his mother, says he might have died from a bubble of air in his heart, although tests were not performed to confirm it and the surgery records indicate that doctors suspected a blood clot.

His official cause of death was "complications of blast injuries."

Doctors say that determining the precise cause of blood clots is rarely possible, making it difficult to establish definitively whether Factor VII is responsible for later complications. And military doctors caution against drawing any conclusions from individual cases.

"A year ago we had a 25-year-old patient, a burn patient, who had a heart attack, and he didn't get Factor VII," said Holcomb, considered one of the world's authorities on the use of Factor VII in trauma cases. "There are lots of complications occurring in this group of significantly injured young people. They have devastating injuries."

But doctors also said that some of the blood clots they have seen in war casualties from Iraq were likely side effects of Factor VII.