Ordinarily, a new medical treatment is approved only after at least one large-scale, randomized controlled trial shows it to be effective without being harmful. Yet at times — and notably during epidemics — evidence can mount that a treatment is safe and effective before we have that final evidence. We think this is the case with Covid-19 and convalescent plasma.
On Sunday, the Food and Drug administration gave a limited, emergency authorization for the inpatient use of convalescent plasma, which is donated by people who have recovered from Covid-19. The decision set off a media firestorm, with politicians, public-health officials and medical professionals insisting that the Donald Trump administration was playing politics with science.
But as two authors of a study the FDA relied on heavily in making its decision, we believe that there are several convincing reasons that going ahead now with emergency use of the treatment is the right way to save lives until the large randomized trials now being mounted can be completed.
First, human plasma is used as a treatment all the time in clinical medicine. Plasma is obtained from donors and distributed by blood banks, and about two million units are used annually in the U.S. Convalescent plasma for patients with Covid-19 differs from ordinary plasma only in that it comes from individuals who have recovered from the disease, and is therefore expected to be rich in antibodies to the SARS-CoV-2 virus.
Second, the largest study to date, done by the Mayo Clinic and colleagues, has produced firm evidence based on the first 20,000 treated patients, that acute side effects were extremely rare. Covid-19 plasma has proved as safe as regular plasma.
Third, an especially a strong piece of evidence has recently emerged from the most recent Mayo Clinic study, and it was this new finding that compelled the FDA to act. Blood banks have a policy of retaining a small amount of plasma or serum from each unit they provide to hospitals, to allow testing of the sample in case an unexpected infection or other complication is found in the recipient. This practice permitted researchers to measure the amount of antibody provided to several thousand transfused Covid-19 patients, and it showed that mortality in recipients was directly and strongly related to the amount of antibody in the transfused plasma.
This finding of a clear dose-response relationship is as close to the result of a randomized trial as could be designed. The amount of antibody in plasma was unknown to anyone when it was provided to the patient. In effect, patients were assigned, in a process akin to randomization, to different levels of antibody, and the results showed that the active principle in convalescent plasma – the antibody level – was effective in reducing mortality. (We do not yet know whether recipients of plasma with lower levels of antibody benefited or might still have fared better than if they had received none.)
This finding was posted on Aug. 12 on the MedRxiv server, which publishes preliminary reports of work that has not been certified by peer review. (It had nearly 40 co-authors, including ourselves.) We know that scientists at the FDA’s division of biologics reviewed this data — and additional data requested from Mayo — before last weekend’s approval announcement. We find no suggestion of unreasonable delay or unreasonable haste in the FDA action. The political overlay that science must sometimes endure can be very misleading.
The Mayo findings — along with two randomized controlled trials that showed reduced mortality with convalescent plasma but below the threshold of statistical significance — point to the effectiveness of convalescent plasma when used appropriately. A meta-analysis of these smaller studies found that plasma administration was associated with a 57% reduction in mortality.
The efficacy of convalescent plasma treatments can also be seen in the long history of successful treatment of infections ranging from meningitis to diphtheria to pneumonia with convalescent plasma. We also know that human convalescent plasma is effective in animals infected with the Covid-19 coronavirus, and that it contains antibodies that neutralize the virus in the laboratory.
More than 1,000 Americans are dying daily of Covid-19. Is it wiser to wait until we have one or more large, properly controlled randomized trials of convalescent plasma before using it as treatment? Not when we can allow broader access for hospitalized patients to a treatment that is demonstrably safe and that has an abundance of evidence of effectiveness.
The FDA decision also addresses a health equity issue. Plasma was already available in the U.S. for some people in hospitals with sufficient resources to navigate the prior access rules, which required physicians to fill in the paperwork needed to secure it for their patients and then continued monitoring. Under the new FDA rules, dispensing plasma is easier for understaffed hospitals caring for underserved populations to get plasma to their patients.
We agree that randomized controlled trials are the gold standard when it comes to the evaluation of disease. But we cannot always wait for gold. And convalescent plasma appears to be the silver lining in the gloomy cloud of Covid-19. – Bloomberg