Fatal in about 90% of cases, the malady had been documented in leukemia and other cancer patients who took Fludara. For that reason, the risk of graft-versus-host disease was noted in the product labeling -- as was a warning about irradiating transfusions as a prevention.
But the NIH doctors did not specify that transfusions should be irradiated for patients in the lupus study. In an interview, Dr. John H. Klippel, then the institute's clinical director, said he could not recall whether he or his colleagues took stock of the label warning.
In Britain, authorities were more cautious, recommending that blood transfusions for all patients taking Fludara be irradiated. The British recommendations were described in 1996 in The Lancet, a medical journal with an international circulation.
Two weeks after being admitted to the NIH Clinical Center, 42-year-old Jamie Ann Jackson died.
"Steve Katz was notified almost immediately," Klippel said.
Katz's subordinates warned the remaining patients and their personal doctors about the death and, for the first time, advised them to irradiate any transfusions. The FDA was informed.
But the NIH office responsible for conducting an inquiry into research deaths was not promptly notified.
And while Katz's institute stopped enrolling recruits, the treatment of those already in the study continued for nine months after Jackson's death.
After five of the other 12 patients given Fludara experienced abnormal changes in their blood, increasing their risk of infection, the experiment was stopped, 20 months before its scheduled conclusion.
'Absolutely No Role'
While Fludara's use for anything other than leukemia remained experimental, an increasing number of doctors were prescribing it "off-label" for diseases of the immune system, including rheumatoid arthritis.
Yet the NIH was slow in warning them about the lethal, but preventable, problem of graft-versus-host disease.
It was not until October 2000, 16 months after Jackson died, that doctors from the NIH briefly summarized the death in Transfusion, the journal of the American Assn. of Blood Banks.
Meanwhile, three articles written by NIH doctors and published from March 2000 through May 2001 referred to the agency's work with Fludara without mentioning the risk of graft-versus-host disease or the death in their study.
