Dr. Edmund Tramont, chief of the National Institutes of Health's AIDS Division, took responsibility for both decisions. He cited his four decades of medical experience and argued that Africans in the midst of an AIDS crisis deserved some leniency in meeting U.S. safety standards, according to interviews and documents obtained by The Associated Press.
Tramont's staff, including his top deputy, had urged more scrutiny of the Uganda research site to ensure it overcame record-keeping problems, violations of federal patient safety safeguards and other issues that forced a 15-month halt to the research into using nevirapine to prevent African babies from getting AIDS from their mothers.
The Associated Pressthat NIH knew about the problems in early 2002 but did not tell the White House before President Bush launched a plan that summer to spread nevirapine throughout Africa. Now, officials have new concerns the drug may cause long-term resistance in the hundreds of thousands of African patients who received it, foreclosing future treatment options.
"I am not convinced that the site is indeed prepared to become active," Dr. Jonathan Fishbein, an expert NIH hired to improve the agency's research practices, wrote Tramont in July 2003.
Fishbein contended he should be given time to review Uganda's capabilities and safety monitoring before letting the site reopen, or NIH would risk being "toothless" in its new efforts to clean up sloppy research practices. He added that professional safety monitors hired by NIH had reservations about the site.
In 2002, less than a month after Health and Human Services Department sent a nine-page letter to Ugandan officials identifying violations of federal patient protection rules by NIH's research.to push nevirapine across Africa to slow the AIDS epidemic, the
The NIH research "may have represented a failure to minimize risk to the subjects," the Office of Human Research Protections told Ugandan authorities in summer 2002.
Africa accounts for more than two-thirds of the world's AIDS cases, with 27 million infected, and the United States sought to help slow the disease's spread across the continent.
Tramont dismissed the safety monitors' concerns, saying he didn't believe they fully understood AIDS.
"I am convinced that this site is ready to resume given the limitations of doing research in any resource-poor, underdeveloped country," Tramont wrote July 8, 2003, in response to Fishbein.
"I want this restriction lifted ASAP because this site is now the best in Africa run by black Africans and everyone has worked so hard to get it right as evidence by the fact that their lab is now certified," he wrote.
NIH officials acknowledge Tramont rewrote the report and overruled his staff on the reopening, but said he did so because he was more experienced and had an "honest difference of opinion" with his safety experts. They noted he had no financial interest in nevirapine and that the troubled study began well before he joined NIH in 2001.
Those who raised objections "were part of a large team of which Dr. Tramont was the head, and it is important that the people involved in that team should express their opinion and there should be discussion," said Dr. H. Clifford Lane, the NIH's No. 2 infectious disease specialist and one of Tramont's bosses. NIH designated Lane to speak to AP on Tramont's behalf.
"But at the end of the day the final responsibility lies with the head of the team and it is his job to put that together the way he sees it," Lane said.
Lane said an internal NIH review concluded Tramont had not engaged in scientific misconduct. Separately, the National Academy of Sciences continues to investigate whether the Uganda research was valid.
NIH believes it helped save hundreds of thousands of African babies by allowing nevirapine to be used in single doses to block the AIDS virus, Lane said. But he acknowledged the research was imperfect, and NIH now believes nevirapine should no longer be a first choice for newborn protection — if other options exist — because of the newly discovered problems about resistance.
Tramont wrote in 2003 e-mails that he reopened the clinics because he didn't want NIH "perceived as bureaucratic but rather thoughtful and reasonable" and that it was important to encourage Africans' fight against AIDS "especially when the president is about to visit them."
Bush visited the continent a few days after Tramont ordered the clinics reopened.
Tramont's actions, however, drew a blunt reply from his top deputy.
"I think we are cutting off our noses to spite our face here," AIDS Division Deputy Director Jonathan Kagan wrote. "...We should not be motivated by political gains and it's dangerous for you, of all people, to be diminishing the value of our monitors."
Tramont prevailed and the research resumed. A few days later, Tramont sent a note to his staff ordering the end of an 18-month-long debate inside NIH over whether the science from the Uganda trials was valid and safe. That debate began in early 20002 when two audits divulged widespread problems with the research.
The Uganda trial "has been reviewed, re-monitored, debated and scrutinized. To do any more would be beyond reason. It is time to put it behind us and move on," Tramont wrote in a July 13, 2003 e-mail instructing his staff that future issues about the drug be handled directly by his office.
Five months earlier, Tramont surprised one of his own medical officers who had written a report summarizing safety concerns uncovered during a second review of the Uganda trial.
Dr. Betsy Smith's report, finished in January 2003, said the Uganda trial suffered from "incomplete or inadequate safety reporting" and that records on patients were "of poor quality and below expected standards of clinical research."
She strongly urged NIH not to make sweeping conclusions about nevirapine based on the Uganda research. "Safety conclusions from this trial should be very conservative," she wrote.
Behind the scenes, Tramont asked to see Smith's report before it was submitted to medical authorities, including the Food and Drug Administration. "I need to see the primary data — too much riding on this report," Tramont wrote Jan. 23, 2003.
A few weeks later, the safety report was published and sent to FDA without Smith's concerns and with a new conclusion.
The study "has demonstrated the safety of single dose nevirapine for the prevention of maternal to child transmission," Tramont's version concluded. "Although discrepancies were found in the database and some unreported AEs (adverse reactions) were discovered ... these were not clinically important in determining the safety profile."
In disbelief, Tramont's staff began inquiring how Smith's report got changed. An answer came back from the top.
"I wrote it," Tramont responded.
Excerpts showing key differences between an original
safety report written by the National Institutes of Health's
Dr. Betsy Smith concerning research on the AIDS drug
nevirapine and the rewritten version of the report by
AIDS research chief Dr. Edmund Tramont.
Smith's original version:
"Acceptable or required timeframes for reporting SAEs
(severe adverse events) and deaths were not followed."
"The safety reporting quality for the HIVNET012 study does
not meet levels expected in prenatal trials sponsored by
DAIDS (NIH Division of AIDS)."
"The supervision or monitoring of the willing and capable
Ugandan site personnel in all aspects of safety, including
subject information regarding treatment risks,
verification of eligibility criteria for mothers and
infants as well as safety reporting does not appear to
have been in place and raises concerns about the study
"Safety reporting did not follow DAIDS reporting
requirements during the conduct of HIVNET012. Safety
conclusions from this trial should be very conservative."
"Site records for safety monitoring and subject visits
were of poor quality and make safety statements very
difficult from the perspective of a review process."
"Monitoring during the trial for safety and clinical
trial management was not in evidence."
Tramont's final version:
"There was some concern expressed by one of the American
physician monitors about the adequacy of standards of
clinical care in Uganda."
"During the full review of 80 mother-infant charts, the
reporting of AEs (adverse events) was found to be
generally complete. The discrepancies that were found
between the database and the source documentation were
due to some missing information in the adverse event
"The remonitoring of review process undertaken by the
safety review panel has shown that there was a
consistent attempt throughout the study to document
AEs and SAEs as evidenced by the large numbers of
such reports ... and the small numbers of missed events
in the remonitoring process."
"HIVNET 012 has demonstrated the safety of single dose
nevirapine for the prevention of maternal to child
transmission of HIV infections. Although discrepancies
were found in the database and some unreported AEs
were discovered during the remonitoring process, these
were not clinically important in determining the