SEATTLE, WASHINGTON—When French scientists presented the results from an Ebola drug trial at a press conference on Monday, they did so with plenty of caveats, but their message was hopeful: The drug, favipiravir, appeared to lower mortality in people with low and medium-high levels of virus in their blood, the researchers told journalists at the Conference on Retroviruses and Opportunistic Infections (CROI) here.
But when study leader Denis Malvy of the University of Bordeaux in France presented more details of the results at CROI, many colleagues were underwhelmed. Several scientists criticized the evidence as well as the design of the trial, which is ongoing at four clinics in Guinea. “It doesn’t tell us anything,” said epidemiologist Scott Hammer of Columbia University, who chairs the meeting.
In his presentation, Malvy focused on a group of 40 patients who began the trial with lower viral loads than 29 others who clearly did not benefit from favipiravir. As he explained on Monday, only six of those 40 patients died—half of what was expected based on similar patients treated at the same clinics over the past 3 months. What’s more, after 4 days of starting treatment on favipiravir, an influenza drug, 51% of these patients had such low levels of the virus in their blood that the standard test could no longer detect it. “There was a signal that monotherapy with favipiravir may decrease viral load,” Malvy said.
Malvy was careful to emphasize the limitations of what he called a “noncomparative, proof-of-concept” study. “The goal of this type of trial is to provide evidence that will allow better design of the trials,” Malvy said. “This type of trial does not provide high level of proof and does not definitively answer a question.” Studies now taking place in France with blood samples from the same patients will more closely analyze changes in Ebola virus levels each day they received the treatment. Researchers will also for the first time assess whether the drug has an effect in monkeys, a widely used animal model for Ebola. And the ongoing study may yield more data.
But Hammer said the trial has "many flaws in it, from study design to the interpretability of the data.” The immune system naturally drives down viral levels, making it difficult to separate the effects of the drug, he noted. Because the study did not compare treated with untreated people who sought care at the same time—it used what’s called a “historical” control—Hammer said he could not determine whether those drops in viral levels or the putative survival benefit are linked to favipiravir.
Since the outbreak exploded in August, researchers have vigorously debated how to set up Ebola drug and vaccine studies, with some advocating randomized controlled studies, in which one group of patients doesn't receive the drug, as the quickest way to get results. Clifford Lane, a deputy director at the U.S. National Institute of Allergy and Infectious Diseases, made that point in his own presentation at the meeting, although he didn't single out the favipiravir study for criticism. “As uncontrolled, so-called compassionate studies, fail to reach clear conclusions, they muddy the waters to the point that it becomes increasingly difficult to carry out the appropriate studies," Lane cautioned.
Together with Liberian officials, Lane is poised to launch a randomized study of an antibody cocktail called ZMapp; that study does include a control group in which patients will receive standard care but not the drug. Lane says Guinea and Sierra Leone have so far refused to join the study because of the ethical concerns about not giving the treatment to all participants. (Sierra Leone has shown some support for the design, however, and discussions are continuing.)
Jean-François Delfraissy, who is coordinating France’s response to Ebola and also heads the country’s HIV/AIDS research program, said he disagreed with Lane’s argument—and opposes the ZMapp trials as designed. “I’ve been a principal investigator for a long time with randomized, controlled trials, and at the beginning of combination therapy for HIV, they were not needed to show the drugs were useful,” said Delfraissy, who co-signed an editorial in The Lancet that opposed this design for Ebola treatments. “It’s difficult to have a randomized trial when mortality is 50% or 60%.”
Delfraissy does not rule out supporting the ZMapp trial—with a modification. “It’s an open question,” he said, noting that he had discussions with Lane here. Specifically, Delfraissy says if the monkey studies and the finer analysis of patient samples have positive results, perhaps the standard of care should include favipiravir as part of the control arm in studies of ZMapp and other experimental treatments.
Hammer says he understands Delfraissy’s perspective but says randomized, controlled trials did play an important role in the history of HIV drug development. He points to a controversial study that in 1997—a year after so-called “triple therapy” proved its value—found that three antiretrovirals were better than two. “We proved it,” Hammer said. “And there was no BS about it later.”
A new report on the ethics of Ebola, released today by the U.S. Presidential Commission for the Study of Bioethical Issues, takes a nuanced view of the questions about study design; the committee doesn't rule out placebo-controlled studies but allows room for other designs as well. "Trial designs should be methodologically rigorous and capable of generating results that are clearly interpretable, acceptable to the host communities and, to the extent possible, minimize delays to completing the research," the report says. "Properly designed placebo-controlled trials can meet these conditions, and innovative designs, such as adaptive randomization, ought to be considered as a means of addressing these research goals."
*The Ebola Files: Science and Science Translational Medicine have made a collection of research and news articles on the Ebola virus and the current outbreak freely available to researchers and the general public.