Colorized transmission electron micrograph of the Ebola virus viron.
In a recent “Perspective” article in The New England Journal of Medicine, three physicians (Drs. Cox, Borio, and Temple) make a strong case for not letting the rush to save Ebola patients tempt us to deviate from good science and skip the randomized controlled trial (RCT). Their arguments cut to the essence of the scientific approach to medicine, and they deserve careful consideration.
Ebola is uniquely scary
Ebola is the kind of threat that really gets our attention. The virus was first identified in 1976, and prior to 2013 there were several small outbreaks in Africa with death rates as high as 90%. This time the death rates are lower, but the numbers are much greater. It has spread to several African countries, and a few cases have even reached the US and Europe due to infected travelers and health care workers. We face a risk that Ebola may become endemic, smoldering along as a constant presence in Africa.
There is no known effective treatment. Fear of Ebola has sparked bizarre conspiracy theories and claims of “natural” cures and prevention kits from homeopaths, alternative medicine advocates like Mercola, and purveyors of remedies like colloidal silver and essential oils. These have been covered on SBM here, here, and here.
Treatment approaches
The current approach consists of early identification, isolation, rigorous infection control protocols with space-suit-like personal protective equipment for health care providers, supportive care with intravenous fluids, and conventional treatments for symptoms and complications. The death rate is lower where good medical care is readily available. As of 31 December 2014, the WHO announced there had been 20,206 reported cases with 7,905 reported deaths, for an overall fatality rate of 39%. But the reported case fatality rate in the three intense-transmission countries among all cases for whom a definitive outcome is known is 71%, and the numbers are uncertain because under-reporting continues to be a challenge.
A number of experimental treatments have been tried or are being considered. Several patients have been given serum transfusions from recovered patients to passively supply them with antibodies to the virus. ZMapp, a combination of monoclonal antibodies, has been used to treat 7 individuals; some of them recovered, some did not. Several anti-viral drugs are under development, and certain selective estrogen receptor molecules and ion channel blockers have also shown promise in vitro.
In the face of such a deadly threat, it seems like we can’t afford to sit back and wait for the results of randomized controlled trials (RCTs) published in peer-reviewed journals. We are tempted to try any reasonable-sounding experimental treatment before the results of clinical trials are in. Some of these experimental treatments might be unsafe and ineffective, but we know Ebola is definitely unsafe and deadly. The doctors who gave their critically-ill patients serum transfusions or ZMapp did exactly what most of us would have done in the same circumstances. But was it really the best thing to do?
The need for RCTs
If we give patients ZMapp without a control group and some of them recover, how are we to know whether ZMapp contributed to their recovery? Some have argued that we wouldn’t want to deny treatment to a control group, and that we can compare the recovery rate with ZMapp to the historical recovery rate without it. That approach is particularly problematic with Ebola, since the historical recovery rate is so variable and so dependent on differences in supportive treatment and other confounding factors. As Cox, et al. point out, “the historical case fatality rates are irrelevant if current study patients receive better supportive care.”
They argue that allowing patients access to investigational drugs outside of properly-designed randomized controlled trials can have tragic consequences. Using no controls or only historical controls might lead to rejection of treatments that have real but modest benefits, or it might lead to widespread adoption of a treatment that is useless or even harmful. And there are ethical considerations. Limited supplies of investigational drugs have been given to favored Western health care workers instead of being equitably distributed to all candidates.
RCTs tend to be slow and cumbersome, but there are ways to streamline the process. They explain:
…advances in trial design can and should be incorporated into Ebola RCTs. For example, such trials should include ongoing monitoring of results (e.g., group-sequential designs), adaptive elements, and other trial efficiencies to reduce the time required to identify an effective treatment, particularly a very effective treatment. If one investigational drug clearly shows benefit, trials should incorporate it into the new standard of care for all treatment groups thereafter. Then a regimen adding a different investigational therapy to the new standard of care could be compared with the new standard of care alone. If multiple investigational drugs are simultaneously available for clinical testing, an RCT could include more than one drug and a shared control group. Trials could be designed to assess effects on survival (recovery from disease) as the most important and measurable end point…
Scientists at the National Institutes of Health, in collaboration with the Food and Drug Administration, the Biomedical Advanced Research and Development Authority, the Department of Defense, and clinicians caring for patients with EVD in the United States, are leading efforts to develop and implement such trials.
This is a principle that applies to all new treatments in mainstream medicine as well as to the treatments offered by so-called “alternative” medicine and “integrative” medicine. A single anecdote about a single patient tells us nothing, and an uncontrolled study is little better than a collection of anecdotes or testimonials. Anything can seem to work in an uncontrolled study; we have seen many examples of studies with apparently positive results even for things as ridiculous as homeopathy, therapeutic touch, and retroactive prayer.
It has been argued that patients should have ready access to untested and experimental treatments, and that they have the right to try anything they want, especially when their life is at stake and there is no treatment available that has been tested and proven effective. In some cases, trying an untested treatment might very well help the individual patient; but if the treatment is really effective, providing it outside of controlled trials tends only to slow the progress of science and delay the day when that treatment will be proven effective and added to the armamentarium of conventional medicine so that it can benefit all patients. We need to determine whether a treatment works or doesn’t work, and there is only one way to do that: using scientific methodology.
Conclusion
From bloodletting for fevers to internal mammary artery ligation surgery for heart disease, the history of medicine is littered with treatments that were believed to work until they were discredited by rigorously-controlled studies. Randomized controlled trials are the most reliable way to determine whether any new treatment is safe and effective. It is tempting to bypass them and offer experimental treatments to desperately-ill Ebola patients outside of a properly designed RCT. We would be wise to resist that temptation and to insist on enrolling patients in a proper clinical study.
Please don’t misunderstand. I am not saying we shouldn’t use untested treatments in a last-resort situation. I am not saying that access to experimental treatments should be restricted. I am only saying that when experimental treatments are used, they should be used as part of a well-designed clinical trial so that we can learn something from the experience. I recognize that enrollment in a clinical trial is not always feasible, but it is a worthy goal. As a society, we should aspire to take whatever actions are necessary to develop and fund such trials for as many patients as possible.