Statement Concerning an Ethical NEJM Study

New England Journal of Medicine Study Proves Mother-Infant HIV Prevention Trials Could Have Been Done Without Placebo

October 4, 2000

The New England Journal of Medicine is publishing in its October 5, 2000 issue a landmark study from Thailand on the use of the drug AZT to prevent HIV transmission from mother to infant. Unlike more than a dozen other studies that provided at least some pregnant women with either placebos or treatments not proven to be effective, this study provided AZT to all pregnant women.

Public Citizen criticized the design of the unethical studies in a September 18, 1997 Journal article. Marcia Angell, then Executive Editor of the Journal, supported Public Citizen in an accompanying editorial that also attacked the use of placebos in these studies.

"This thoughtfully designed study is superior to previous studies in both its ethics and its scientific design. The study proves that science can be conducted both rigorously and ethically," said Dr. Peter Lurie, deputy director of Public Citizen's Health Research Group.

The present study, authored by Dr. Marc Lallemant of the Harvard School of Public Health and his colleagues, randomly assigned 1,437 HIV-positive pregnant women to groups in which the women received either long (12 weeks) or short (5 weeks) durations of AZT before delivery and the infants received either long (6 weeks) or short (3 days) durations of AZT. There were thus four groups: long-long (similar to a regimen called ACTG 076 proved effective in the U.S. and France in 1994); long-short; short-long; and short-short. All women received AZT during labor, usually by mouth.

The HIV transmission rate in the short-short arm (10.5%) was so high that the researchers prematurely terminated that arm. Even so, the short-short arm was so much better than the placebo arms in all of the unethical studies (and better than the rates of transmission in other non-treatment studies conducted before AZT was proved effective for this purpose) that there was no difficulty concluding that short-short was better than nothing. The placebo in the other studies was therefore unnecessary and dangerous.

The transmission rates for the long-long (6.5%), long-short (4.7%) and short-long (8.6%) were statistically indistinguishable. The study provides clinicians with much better guidance as to which part of the AZT regimen is critical than do most of the other placebo-controlled trials.

"Had the researchers in the unethical studies designed theirs like the present study, hundreds of infants would have been saved because the mothers would have received a drug instead of placebo," said Dr. Sidney M. Wolfe, director of Public Citizen's Health Research Group.

Public Citizen also noted that the present study is being published after the first developing country placebo-controlled trial (in Thailand) for reasons unrelated to the lack of a placebo. The study began recruiting patients 13 months after the placebo-controlled trial in Thailand. Furthermore, it had four study arms, compared to two in the placebo-controlled trial, requiring more patients and hence more time to recruit.

Coincidentally, the World Medical Association is meeting this week in Edinburgh to discuss possible revisions to the Declaration of Helsinki. At issue, in particular, is language that could permit substandard therapies (like placebos) to be provided by researchers in poor countries.

"A grassroots effort from around the world appears to have succeeded in repelling the forces in the research industry that sought to dilute the Declaration's protections for poor patients," said Dr. Lurie. "Now we can focus on actually getting AZT and other drugs to the people who need them."