Convalescent Plasma: What We Still Need to Know
Convalescent plasma—the liquid portion of the blood obtained from COVID-19 survivors—is rich in antibodies that can fight the novel coronavirus.
Soon after the COVID-19 pandemic began in the United States, researchers started collecting plasma from survivors to transfuse into patients who were sick with the disease, in the hope that the antibodies would help those patients recover. But to get the chance to be treated with convalescent plasma, most patients had to enroll in a clinical trial.
In late August, the U.S. Food and Drug Administration granted an Emergency Use Authorization for convalescent plasma and now any hospitalized patient with COVID-19 can receive it with few, if any, restrictions.
At Columbia, transfusion medicine expert Steven Spitalnik, MD, professor of pathology & cell biology at Vagelos College of Physicians and Surgeons, is participating in a prospective randomized clinical trial and other studies of this treatment, which he says are still vital to conduct.
CUIMC News spoke with Spitalnik and his collaborator James Zimring, MD, PhD, professor of pathology at the University of Virginia School of Medicine, about the therapy.
Why is there such a big interest in convalescent plasma therapy for COVID-19 patients?
JZ: Antibodies are known to play a role in fighting some viral illnesses, and many people who recover from COVID-19 have antibodies against the virus that causes the disease. The antibodies could be used to either prevent illness in people exposed to COVID-19 or treat those who are currently sick with the disease.
It’s a very reasonable therapeutic approach to try; however, convalescent plasma therapy has worked for some types of infection, but not for others, so it must be rigorously tested before we know if it helps, harms, or has no effect.
Why do we need to study convalescent plasma in randomized clinical trials?
SS: Without appropriately designed prospective trials, we do not know that it is effective, nor do we know that it does no harm. Even if it does work, we do not know much should be given, when it should be given, and what levels and types of antibodies are important.
To tell if convalescent plasma is effective, you need two groups of patients where the only difference between them is whether they receive convalescent plasma or placebo. Of course, all patients differ in multiple ways. To account for these differences, patients are randomly assigned to a therapy group or a placebo group. In addition, the nature of the placebo is relevant, because there are other components of plasma, besides antibodies, that may be beneficial or harmful; thus, we use non-COVID-19 plasma as the placebo control. Ideally, neither the patients nor the health care providers know which treatment they will receive to minimize the risk of bias and confounding.
Can’t we just compare the thousands of patients who have already received convalescent plasma with those who didn’t?
SS: The problem is that outside of the controlled conditions of a randomized clinical trial, multiple variables are changing at the same time. These variables could make it look like convalescent plasma is working when it really isn’t. For example, the doctors giving convalescent plasma might be more aggressive than some of their colleagues and may also be giving additional treatments that could affect outcomes, for better or worse.
Conversely, convalescent plasma may appear to have no therapeutic value when it really is working. For example, in the hospital setting, doctors may only transfuse the sickest patients. These patients may do better after getting the therapy than they otherwise would have, but their outcomes may still be worse compared to less sick patients who didn’t get convalescent plasma.
What’s the potential harm of authorizing use of a treatment before clinical trials have been completed?
JZ: Time and again, the history of medicine offers examples of mistakes in disease treatment because of the lack of randomized controlled trials.
Perhaps the most famous example was the use of hormone replacement therapy to decrease the risk of heart disease in postmenopausal women. It made sense that if physicians replaced the hormones women lose during menopause, the risk of heart disease should decrease to premenopausal levels. Indeed, women treated with hormone replacement therapy seemed to have lower rates of heart disease, and the treatment became widely used.
Out of an abundance of caution, large randomized trials were performed anyway, and the finding was shocking: Hormone replacement therapy was found to either have no effect or increase risk of heart disease. How could this be?
The answer is that, in an uncontrolled setting, women who used hormone replacement therapy tended to be more proactive in their health care and engaged in other lifestyle practices that decreased risk of heart disease. So the benefit was derived from these other behaviors, not the hormone replacement therapy. Indeed, the benefits from these behaviors offset the detrimental effects of hormone replacement therapy, hiding the harm that was being done.
Without the rigor of randomized controlled trials, we would still be giving postmenopausal women a therapy that likely caused them harm.
What if convalescent plasma not only has no benefit, but also causes harm, as was observed with hydroxychloroquine? In that case, not only would we make things worse, but we would never know we had done so.
Why do you think the Emergency Use Authorization will hinder research on convalescent plasma?
SS: If you were in a hospital with COVID-19 and were given a choice between getting FDA-authorized convalescent plasma or enrolling in a trial where you might be randomized to receive either the treatment or a placebo, what would you choose?
Emergency Use Authorization may severely inhibit the ability of rigorous scientific trials to be performed and completed in the United States, because patients will no longer enroll in them in sufficient numbers, if at all. This would prevent us from ever really knowing if convalescent plasma is effective and, if so, determining the best way to use it.
JZ: Among the great benefits that medical research has provided to patients over the last century is the ability to tell if therapies actually work and how to use them if they do. By jumping the gun on prescribing a treatment before we have sufficient information, we are likely to do more harm to patients, and at a massive scale, than will be done by waiting a little longer for ongoing clinical trials to be completed.