Letter in Medical Journal of Australia on Perinatal HIV Transmission Prevention Trials
Letter on Perinatal HIV Transmission Prevention Trials
This letter, by Health Research Group Deputy Director Peter Lurie, M.D., MPH, and Director Sidney Wolfe, M.D., appeared in Medical Journal of Australia on August 2, 1999.
To the Editor: McNeill states that the placebo-controlled trials of shorter courses of zidovudine to prevent perinatal HIV transmission in Africa and Asia “were quicker” than active controlled trials and “will save more lives”.[1] These assertions are incorrect for four reasons.
First, data available in 1993 showed that most perinatal HIV transmission occurs in the last weeks of pregnancy and during delivery, making it very likely that regimens concentrating on these periods would be effective.[2] In addition, data from the original zidovudine study available in 1994 showed no univariate association between duration of zidovudine treatment (mean 7 weeks vs. mean 17 weeks) and its effect in reducing perinatal HIV transmission compared with placebo.[3]
Second, the speed of completing a clinical trial is related to the sample size required. But sample size requirements for active-control trials are very often similar to those needed for placebo-controlled ones.[4] In the specific case of the perinatal trials, we have previously demonstrated that the sample size required for an “equivalency” study in which short courses of zidovudine are compared with the standard, longer zidovudine regimen requires 620 subjects compared with 500 subjects for a placebo-controlled trial,[3] an insignificant difference. More active recruiting or opening additional study sites can readily compensate for this.
Third, the National Institutes of Health (NIH) did approve a Harvard University equivalency study of perinatal HIV prevention in Thailand even as the placebo-controlled trials were being proposed. This study had four arms and a sample size requirement of 1554, actually less than the only other four-arm study, conducted by the United Nations Programme on HIV/AIDS, which had a placebo group and a target sample size of 1900. That equivalency study might have been finished today had not reviewers at the NIH insisted upon a placebo control. This prompted multiple letters back and forth between the agency and the researchers before the NIH capitulated.[3] By then the study had been delayed for several years.
Fourth, the vagaries of health policy making are such that minor differences in sample size requirements have an insignificant impact upon actual delivery of services. This is best illustrated in South Africa, where the strikingly positive results of a placebo-controlled trial (all of the three other placebo-controlled antiretroviral drug trials in developing countries have also been positive), to which South Africa contributed 52% of the subjects, have failed to convince the national government to fund even a pilot program using zidovudine.[5]
The greatest contributor to unnecessary infant HIV infection in developing countries comes not from subtle differences in study design or sample size, but from the failure of local governments and sponsoring countries to act five years ago to implement short-course regimens.
[1] McNeill PM Should research ethics change at the border? [editorial] Med J Aust 1998; 169 509-510
[2] Rouzioux C, Costagliola D, Burgard M, et al Timing of mother-to-child HIV-1 transmission depends on maternal status AIDS 1993; 7 Suppl 2: S49-S52
[3] Lurie P, Wolfe SM Unethical trials of interventions to reduce perinatal transmission of the human immunodeficiency virus in developing countries New Engl J Med 1997: 337 853-856
[4] Freedman B, Weijer C, Glass KC. Placebo Orthodoxy in clinical research. I. Empirical and methodological myths. J Law Med Ethics 1996; 24 243-251
[5] South Africa: setting the wrong example [editorial] Nature Med 1999; 5:1