It was early April and I was working an overnight shift in the intensive care unit when I found myself faced with a decision. My patient, a relatively young man with severe Covid-19 and respiratory failure, was spiking high fevers, his blood pressure dropping.
I had to decide whether or not to give him an immunosuppressive drug called tocilizumab. If my patient was worsening because of the inflammatory cytokine storm that accompanies some cases of coronavirus, then the drug might help. But if he actually had a bacterial infection in addition to the virus, suppressing his immune system could be risky.
Facing a critically ill patient and without any clear evidence to guide me, I felt a sense of desperation. Let’s give the med, I decided. I had given it before, and those patients had done well. After all, I felt, we had to do something.
I thought back to that moment recently, when the Food and Drug Administration granted emergency use authorization for convalescent plasma despite inadequate evidence. Back in April when I was deciding whether to administer tocilizumab to my patient, we were all making decisions in the dark, driven by fear and anecdote rather than data. That was excusable in the early weeks of the surge in cases. But now, months later, we are in much the same place. And that is not excusable.
As a doctor at the bedside, I need high-quality evidence to be able to make the right decisions for my patients. Without it, the public’s trust in science and our lives are at stake.
The history of medicine is rife with promising therapies that have later proven to be ineffective or, even worse, harmful. When I was a doctor in training first working in the intensive care unit, I learned to use a drug called activated drotrecogin alfa (sold as Xigris), which was touted as a lifesaver for patients suffering from severe sepsis. I distinctly remember giving it to patients who subsequently improved, seemingly pulled back from the brink.
We are wired to infer cause, and so I congratulated myself for my decision and used this drug whenever I could. But when it was studied in a rigorously designed clinical trial, researchers found that it was no better than a placebo, and it was taken off the market.
That’s why we have to do clinical trials. They are the gold standard to test a treatment’s efficacy but have been scarce in the pandemic. In these trials, investigators randomly assign patients to receive either the drug that’s being studied or — as a control — a placebo or standard care.
This relies on a concept called clinical equipoise, which means that the medical community at large is uncertain about whether a new treatment is actually better or worse than standard care. If we knew beyond doubt that something was lifesaving — like using a parachute when jumping out of an airplane in flight — we couldn’t ethically test it against a placebo on our patients.
Additionally, patients who consent to be part of a randomized trial must be willing to risk a 50 percent chance of receiving the placebo. The process becomes far more challenging, then, when the drug that is being studied is broadly available outside the trial framework.
This is the case with convalescent plasma. The idea behind it is attractively simple: Harvest blood plasma from patients who have recovered from Covid-19, test it to make sure it has antibodies to the virus and then inject it into patients with the virus to help them fight it until their own antibodies kick in.
Early in the pandemic, when the buzz about convalescent plasma was first building here in the United States, I cared for a patient with Covid-19 who was not getting better. And as the threat of her death loomed, her family grew frantic. They pleaded for us to give her convalescent plasma and promised to do whatever it took. I, too, felt the uncomfortable need to do something more than manage volume status and tweak ventilator settings.
Had convalescent plasma been available at my hospital, I would have undoubtedly tried to treat my patient with it even in the absence of compelling data. But it was not. (It is now, though within the constraints of a placebo-controlled trial). She ultimately recovered nonetheless.
Since that time, more than 70,000 patients in the country have received convalescent plasma. But we still don’t know if it helps.
The data cited in the news conference announcing the emergency approval come from a large continuing study of more than 35,000 patients, and though the benefit was grossly inflated, the true results do suggest that convalescent plasma might have been associated with fewer deaths in patients who got it early on in their hospital course. But there are major drawbacks to making any conclusions here.
The study, which has not been peer reviewed or published in a medical journal, is part of an “expanded access program.” This is not a randomized trial but instead a way to get a potentially promising treatment to patients while tracking how they do.
As a result, there was no control group for comparison, which makes it impossible to state with certainty that convalescent plasma is really the reason for an observed decrease in mortality. Maybe convalescent plasma has a small or moderate benefit. But it might also be neutral or even harmful. The question remains open.
And now we might never be able to answer it. Because if convalescent plasma is widely available and hailed as a major breakthrough, who is going to sign up for a trial where a roll of the dice might leave the patient with a placebo?
In the haste to do something and the inability to sit with uncertainty, we might have deprived ourselves of the very answers that we need. This will be one of the many sad legacies of our response to this pandemic.
As a nation, we have battled the virus for nearly six months, with more than six million people infected. Yet only a small minority of our patients with Covid-19 have been enrolled in clinical trials. In contrast to the national clinical trials in Britain that have brought us valuable preliminary data on dexamethasone and hydroxychloroquine, the fractured nature of our health care system means that multiple different institutions work in silos trying to ramp up their own trials and competing for patients rather than collaborating.
Even now, treatment protocols vary across the country. Well-intentioned clinicians continue off-label use of a panoply of therapies while defending their decisions with an almost religious fervor.
This is not just about the coronavirus. It is about the extent to which we are committed to science and to each other. As we look toward a vaccine, which will rely on the trust of the public and the medical community for widespread dissemination, this would have been the time to get it right.
I know what it is to stand with a person who might be dying in front of you, to hear your own voice and those of your patient’s loved ones in your head, asking you whether you’re sure there isn’t anything more that you can try.
I did give the drug to my patient that night in April before learning that his fever was more likely due to a new bacterial pneumonia. For days I worried that I had harmed him and wished that I had something other than desperation to go on when making my decision. Ultimately, with treatment of his pneumonia and meticulous critical care, he recovered.
Looking back, I have no idea if the tocilizumab hurt, helped or had no effect at all. There are still no rigorous trials to tell me. But I do know that the next time I am faced with a similar decision at the bedside, I want to be able to take a breath. And to turn to the science.
Daniela J. Lamas is a critical care doctor at Brigham and Women’s Hospital in Boston.
The New York Times
