| ||||||
 |  |
|  | | | |
| Farber | The 1986 Phase II trial that preceded the FDA’s unprecedented rapid approval of AZT was presented as a double-blind, placebo-controlled study, though it was anything but that. As became clear afterward through the efforts of a few journalists, as well as the testimony of participants, the trial was “unblinded” almost immediately because of the severe toxicity of the drug. Members of the control group began to acquire AZT independently or from other study participants, and eventually the study was aborted and everyone was put on the drug. |
| Gallo | Farber states that the “AZT study was unblinded almost immediately because of the severe toxicity of the drug. Members of the control group began to acquire AZT independently or from other study participants.” Farber cannot have it both ways. If the BW 002 study became unblinded because of AZT's toxicity, then control group members would surely not have wished to acquire AZT. If the study became unblinded because AZT tasted differently to placebo, then perhaps control group members might have tried to acquire it. But here again, Farber makes a series of old AIDS denialist allegations that the results of BW 002 are invalid because of irregularities in the trial. In effect, Farber asks readers to take the side of a journalist who does not believe that HIV causes AIDS, John Lauritsen, against the considerably more expert opinion of the FDA panel that approved AZT. A number of points need to be made about this: (1) All the trial participants were symptomatic of AIDS or what was called AIDS Related Complex at the time. One out of 145 AZT recipients died on the trial. Nineteen out of 137 placebo recipients died. Furthermore the AZT recipients had fewer opportunistic infections and scored higher on quality of life measurements. This cannot be explained by chance and demonstrated the efficacy of AZT. Hence the FDA registered it. (2) If as is alleged by the AIDS denialists some subjects became unblinded with the consequence that placebo subjects took AZT, then the results of the trial actually underestimate the efficacy of AZT and the AIDS denialist case is hoisted by its own petard. This is because if AZT was more dangerous than placebo, then there should have been more than just one death on the AZT arm. If the allegation of unblinding is true, then the only logical conclusion is that the number of placebo deaths was fewer than should have been the case, because some of the placebo subjects were given extra life-expectancy by taking AZT. There is simply no logical way for AIDS denialists to explain the massive difference in life-expectancy between the two arms. (3) BW 002 was not the only placebo-controlled study that demonstrated AZT's efficacy. A placebo controlled trial known as ACTG 016 showed that symptomatic patients with CD4 counts between 200 and 500 were less likely to progress to AIDS. No difference in disease progression was seen in patients with CD4 counts greater than 50055 (4) Fifteen AZT versus placebo studies have been conducted. Not one shows any evidence to support AIDS denialist arguments that AZT causes AIDS or that its risks outweigh its benefits. 56 (5) Several uncontrolled studies have shown that AZT increases life-expectancy in symptomatic HIV patients. 57 (6) The BW 002 trial that Farber refers to in the main text involved AZT use as monotherapy. As is now well understood. HIV mutates rapidly resulting in selection for strains of the virus that are resistant to a single drug. Indeed, if AZT was not effective, HIV would not need to mutate to escape it. The short-term benefits demonstrated in the first, placebo-controlled AZT study led to the demand that subsequent trials of potential antiretroviral drugs in patients who had progressed to AIDS did not use a placebo control, but rather employed AZT. Consequently, subsequent studies demonstrated improved survival in individuals receiving dual drug therapy compared to AZT. (7) Farber makes no mention of the fact that numerous ARV trials have demonstrated that they reduce morbidity and mortality. A meta-analysis of ARV clinical trials found the following: One ARV reduces progression to AIDS or death by 30% against placebo. Two ARVs reduce progression to AIDS or death by 40% against one ARV. Three ARVs reduce progression to AIDS or death by 40% against two ARVs. (Jordan et al. BMJ. 2002 March 30; 324(7340): 757.) If the risks of ARVs outweigh their benefits, why does using more of them result in less mortality and morbidity? (8) A recent ARV study by the NIH, the largest ever conducted, found that the continuous use of ARVs resulted in half the rate of disease progression and death than occurred when treatment was interrupted. 58 If the risks of ARVs outweigh their benefits, why does taking them continuously result in less mortality and morbidity than taking them occasionally? (9) As cited above (see note regarding page 38 column 3), numerous cohort analyses from around the world, both in developing and wealthy countries, demonstrate that ARVs are prolonging and improving life substantially. More examples of the efficacy of ARVs from different cohorts are being published regularly. There is much more but the above should be sufficient to demonstrate that Farber’s arguments are without merit. None of this should imply that ARVs are not associated with side-effects, which in rare circumstances are fatal. But the evidence is beyond doubt that their benefits outweigh their risks. |
| RA | The Gallo document notes correctly that when a trial becomes unblinded the action of participants may be to blur the difference between the trial arms by sharing medications or by purchasing the active drug if the participant is on placebo. On the other hand, the opposite effect may occur if the placebo participants have a powerful feeling that they will die because they are being denied the active medication. At least as important, however, are the actions of the researchers. If they believe that those on the active drug (e.g. AZT) are more likely to live they may consciously or unconsciously treat them in different ways. This may be because they truly believed that AZT was going to save their life and that those on placebo were doomed. Or it could be that they saw the side effects of AZT and undertook desparate measures, such as blood transfusions, to try to avoid facing the reality that the drug was dangerously toxic. All that can be said for certain is that unblinding a trial invalidates it. We describe some of the follow-up evidence that illustrates that the BW 002 trial was not a success in rethinkingaids.com/GalloRebuttal/Farber-Gallo-32.html. While the Gallo document claims that “Not one [clinical trial of AZT] shows any evidence to support AIDS denialist arguments that AZT causes AIDS or that its risks outweigh its benefits.” they are disagreeing with the manufacturer of the drug which notes in their product monograph that “It was often difficult to distinguish adverse events possibly associated with administration of RETROVIR® (AZT™) from underlying signs of HIV disease or intercurrent illnesses”. The Jordan meta-analysis is totally reliant on the AZT trials that are the only placebo-controlled trials of AIDS drugs. And, of all the AZT trials they quote, the Concorde trial is the biggest trial, and the one that adds the most statistical validity. When the marginally positive results of the Concorde trial are mixed with the highly positive results of several much smaller trials, a degree of statistical significance is obtained. The problem with this analysis is that the Concorde trial was not a true placebo-controlled trial. It was a trial of Immediate AZT (at beginning of trial) versus Deferred AZT (when AIDS was diagnosed). So the 'placebo' arm contained many people taking AZT. Of the 172 trial participants who died during the trial, 96 died in the Immediate arm and only 76 in the Deferred arm. What is less frequently reported is that, because those in the Deferred arm were eligible to take AZT, 169 died while taking AZT but only 3 while still taking placebo. Our detailed discussion of the Jordan and Concorde studies is described at rethinkingaids.com/GalloRebuttal/Farber-Gallo-03.html. |